Get documents about "Issues in CDMA
Document Sample
Montreal, QC, Canada f04a7860-1fbf-4a5e-a979-a337dbd4afa8.doc
Title: Joint Proposal for 3GPP2 Physical Layer for FDD Spectra
Source: China Unicom, Huawei Technologies, KDDI, LG Electronics, Lucent
Technologies, Motorola, Nortel Networks, QUALCOMM Incorporated, RITT,
Samsung Electronics, ZTE Corporation
Abstract: This contribution presents a framework proposal for the Phase 2 evolution of
the 3GPP2 air interface for FDD spectra. It describes the Physical Layer with
some MAC Layer constructs.
Date: July 31, 2006
Recommendation: Review and adopt
Notice
Contributors grant a free, irrevocable license to 3GPP2 and its Organization Partners to incorporate
text or other copyrightable material contained in the contribution and any modifications thereof in
the creation of 3GPP2 publications; to copyright and sell in Organizational Partner‟s name any
Organizational Partner‟s standards publication even though it may include portions of the
contribution; and at the Organization Partner‟s sole discretion to permit others to reproduce in whole
or in part such contributions or the resulting Organizational Partner‟s standards publication.
Contributors are also willing to grant licenses under such contributor copyrights to third parties on
reasonable, non-discriminatory terms and conditions for purpose of practicing an Organizational
Partner‟s standard which incorporates this contribution.
This document has been prepared by the Contributors to assist the development of specifications by
3GPP2. It is proposed to the Committee as a basis for discussion and is not to be construed as a
binding proposal on Contributors. Contributors specifically reserve the right to amend or modify the
material contained herein and nothing herein shall be construed as conferring or offering licenses or
rights with respect to any intellectual property of Contributors other than provided in the copyright
statement above.
Table of Contents
1 INTRODUCTION.....................................................................................................................................................10
1.1 FORWARD LINK OF HKLLMNQRSUZ FDD ....................................................................................................11
1.1.1 FDD Superframe .........................................................................................................................................11
1.1.2 Rotational OFDM .......................................................................................................................................13
1.1.2.1 Overview of Rotational OFDM................................................................................................................................ 13
1.1.2.2 R-OFDM Transmitter .............................................................................................................................................. 14
1.1.3 Forward Link Resource Channel ................................................................................................................15
1.1.3.1 DRCH ...................................................................................................................................................................... 15
1.1.3.2 BRCH ...................................................................................................................................................................... 16
1.1.4 Multiplexing of Resource Channel ..............................................................................................................17
1.1.4.1 Multiplexing Mode 1 ............................................................................................................................................... 17
1.1.4.2 Multiplexing Mode 2 ............................................................................................................................................... 17
1.1.5 Forward Link Channels ..............................................................................................................................18
1.1.5.1 Forward Link Pilot Channel ..................................................................................................................................... 19
1.1.5.1.1 Forward Common Pilot Channel (F-CPICH) .................................................................................................... 19
1.1.5.1.1.1 MIMO Support with F-CPICH ............................................................................................................... 20
1.1.5.1.2 Forward Dedicated Pilot Channel (F-DPICH) ................................................................................................... 20
1.1.5.1.2.1 MIMO Support with F-DPICH ............................................................................................................... 21
1.1.5.2 F-DCH ..................................................................................................................................................................... 22
1.1.5.3 Forward Link Control Channels .............................................................................................................................. 23
1.1.5.3.1 Forward Link Shared Control Channel (F-SCCH) ............................................................................................ 24
1.1.5.3.1.1 Summary of Forward Link Control Channel Messages ........................................................................ 25
1.1.5.3.1.2 Modulation of Forward Link Control Channel Messages ..................................................................... 25
1.1.5.3.2 Acknowledgement Channel (F-ACKCH) .......................................................................................................... 26
1.1.5.3.3 Reverse Link Power Control Channel (F-PCCH) .............................................................................................. 26
1.1.5.3.4 Reverse Link Pilot Quality Indicator Channel (F-PQICH) ................................................................................ 26
1.1.5.3.5 Fast OSI Channel (F-FOSICH) .......................................................................................................................... 26
1.1.5.3.6 Other Channels .................................................................................................................................................. 27
1.1.6 MIMO Design .............................................................................................................................................27
1.1.6.1 Data Channel Structure ............................................................................................................................................ 27
1.1.6.2 Pilot Structure .......................................................................................................................................................... 27
1.1.6.2.1 DRCH ................................................................................................................................................................ 27
1.1.6.2.2 BRCH ................................................................................................................................................................ 27
1.1.6.3 STTD Mode ............................................................................................................................................................. 27
1.1.6.4 MIMO Modes .......................................................................................................................................................... 28
1.1.6.4.1 Single Codeword (SCW) Design ....................................................................................................................... 29
1.1.6.4.1.1 Rate and Rank Prediction ........................................................................................................................ 29
1.1.6.4.1.2 Transmitter Structure .............................................................................................................................. 29
1.1.6.4.1.3 Example of Receiver Structure ................................................................................................................ 30
1.1.6.4.1.4 Reverse Link Feedback Channels for SCW ............................................................................................ 30
1.1.6.4.2 Multi Codeword (MCW) Design ....................................................................................................................... 30
1.1.6.4.2.1 Rate and Rank Prediction ........................................................................................................................ 31
- 3-
1.1.6.4.2.2 Transmitter Structure .............................................................................................................................. 31
1.1.6.4.2.3 Example of Receiver Structure ................................................................................................................ 31
1.1.6.4.2.4 BL H-ARQ ................................................................................................................................................ 32
1.1.6.4.2.5 Reverse Link Feedback Channels for MCW .......................................................................................... 32
1.1.6.5 Precoding ................................................................................................................................................................. 33
1.1.6.6 Space Division Multiple Access .............................................................................................................................. 33
1.2 REVERSE LINK OF HKLLMNQRSUZ FDD ..................................................................................................... 33
1.2.1 Reverse Link Resource Channel ................................................................................................................. 33
1.2.1.1 Dedicated Pilot Channel .......................................................................................................................................... 34
1.2.1.2 Multiplexing Diversity and Subband Hopping ........................................................................................................ 35
1.2.2 Reverse Link Channels................................................................................................................................ 35
1.2.2.1 R-DCH ..................................................................................................................................................................... 36
1.2.2.2 Reverse Link CDMA Traffic (R-CDCH) ................................................................................................................. 37
1.2.2.3 Multiplexing CDMA and OFDMA Traffic .............................................................................................................. 38
1.2.2.4 Reverse Link Control Channels ............................................................................................................................... 39
1.2.2.4.1 Reverse Acknowledgement Channel (R-ACKCH) ............................................................................................ 40
1.2.2.4.2 Reverse Pilot Channel (R-PICH) ....................................................................................................................... 42
1.2.2.4.3 Channel Quality Indicator Channel (R-CQICH) ................................................................................................ 43
1.2.2.4.4 MIMO Channel Quality Indicator Channel (R-MCQICH) ................................................................................ 43
1.2.2.4.5 Reverse Link Auxiliary Pilot Channel (R-AuxPICH) ........................................................................................ 43
1.2.2.4.6 Request Channel (R-REQCH) ........................................................................................................................... 43
1.2.2.4.7 Beamforming Feedback Channel (R-BFCH) ..................................................................................................... 43
1.2.2.4.8 Reverse Subband Feedback Channel (R-SFCH) ................................................................................................ 43
1.2.2.4.9 Reverse Access Channel (R-ACH) .................................................................................................................... 44
1.3 OPERATIONAL DETAILS OF HKLLMNQRSUZ FDD ...................................................................................... 44
1.3.1 Acquisition .................................................................................................................................................. 44
1.3.1.1 TDM Pilot 1 ............................................................................................................................................................. 45
1.3.1.2 TDM Pilots 2 and 3 .................................................................................................................................................. 46
1.3.1.3 Synchronization Modes............................................................................................................................................ 46
1.3.1.3.1 Semi Synchronous Mode ................................................................................................................................... 46
1.3.1.3.2 Asynchronous Mode .......................................................................................................................................... 47
1.3.1.4 Acquisition Procedure .............................................................................................................................................. 47
1.3.2 Paging ......................................................................................................................................................... 48
1.3.2.1 SFN Operation ......................................................................................................................................................... 48
1.3.3 Access Channel Procedures ........................................................................................................................ 48
1.3.3.1 Access Probe Structure ............................................................................................................................................ 49
1.3.3.2 Access Probes Transmission Procedure ................................................................................................................... 49
1.3.4 Timeline ...................................................................................................................................................... 50
1.3.4.1 Forward Link Timeline ............................................................................................................................................ 50
1.3.4.2 Reverse Link Timeline ............................................................................................................................................. 52
1.3.4.3 Half Duplex Mode ................................................................................................................................................... 54
1.3.5 Resource Management ................................................................................................................................ 55
1.3.6 Scheduling .................................................................................................................................................. 55
1.3.7 Assignment Management ............................................................................................................................ 56
- 4-
1.3.7.1 Forward Link Channelization .................................................................................................................................. 57
1.3.7.1.1 Resource Allocation ........................................................................................................................................... 57
1.3.7.2 Resource Adaptive H-ARQ (RAS H-ARQ) ............................................................................................................. 57
1.3.7.3 Dynamic Resource Sharing between Forward Link Control and Traffic ................................................................. 58
1.3.7.4 Study Items for Assignment ..................................................................................................................................... 60
1.3.8 Group Resource Allocation .........................................................................................................................60
1.3.9 Reverse Link Power Control (PC) ..............................................................................................................64
1.3.9.1 Open Loop Power Adjustments ............................................................................................................................... 65
1.3.9.2 Reverse Link Control Channel Power Control ........................................................................................................ 65
1.3.9.2.1 R-PICH .............................................................................................................................................................. 65
1.3.9.2.2 R-CQICH ........................................................................................................................................................... 65
1.3.9.2.2.1 To Reverse Link Serving Sector .............................................................................................................. 65
1.3.9.2.2.2 To Forward Link Serving Sector............................................................................................................. 66
1.3.9.2.2.3 To Desired Forward Link Serving Sector ............................................................................................... 66
1.3.9.2.2.4 To Other Sectors ....................................................................................................................................... 66
1.3.9.2.3 R-SFCH and R-BFCH ....................................................................................................................................... 66
1.3.9.2.4 R-REQCH .......................................................................................................................................................... 67
1.3.9.2.4.1 To Reverse Link Serving Sector .............................................................................................................. 67
1.3.9.2.4.2 To Desired Reverse Link Serving Sector ................................................................................................ 67
1.3.9.2.5 R-ACKCH ......................................................................................................................................................... 67
1.3.9.3 Reverse Link OFDMA Traffic Channel Power Control ........................................................................................... 67
1.3.9.3.1 RDCH Gain Determination ................................................................................................................................ 70
1.3.9.4 Reverse Link CDMA Traffic PC .............................................................................................................................. 71
1.3.10 Fractional Frequency Reuse .......................................................................................................................72
1.3.11 Subband Scheduling....................................................................................................................................72
1.3.11.1 Local Hopping and Channel Trees ...................................................................................................................... 72
1.3.11.2 Subband Feedback ............................................................................................................................................... 73
1.3.12 Quasi Orthogonal Reverse Link ..................................................................................................................73
1.3.12.1 Random Hopping ................................................................................................................................................ 74
1.3.12.2 Multiplexing Factor Control Through Scheduling .............................................................................................. 74
1.3.12.3 Orthogonal Pilot Multiplexing ............................................................................................................................ 75
1.3.13 Layered Superposed OFDMA (LS-OFDMA) ..............................................................................................76
1.3.14 Handoff Scenarios.......................................................................................................................................76
1.3.14.1 Active Set Management ....................................................................................................................................... 77
1.3.14.2 Forward Link Handoff ......................................................................................................................................... 77
1.3.14.2.1 Softer Handoff Groups ..................................................................................................................................... 78
1.3.14.3 Reverse Link Handoff.......................................................................................................................................... 78
2 EPILOGUE ............................................................................................................................................................... 79
3 REFERENCES..........................................................................................................................................................80
4 APPENDIX ................................................................................................................................................................ 81
4.1 PRECODING AND SDMA CODEBOOKS ............................................................................................................81
4.2 CODING AND MODULATION FOR THE R/F-DCH ............................................................................................81
- 5-
4.2.1 Channel Coding .......................................................................................................................................... 81
4.2.1.1 Packet Splitting ........................................................................................................................................................ 81
4.2.1.2 Convolutional Code ................................................................................................................................................. 81
4.2.1.3 Turbo Code .............................................................................................................................................................. 81
4.2.2 Channel Interleaving .................................................................................................................................. 82
4.2.3 Incremental Redundancy by Puncturing and Repetition ............................................................................ 82
4.2.4 Data Scrambling ......................................................................................................................................... 83
4.2.5 Modulation Formats and Modulation Step-down ....................................................................................... 83
4.3 GCL SEQUENCES .............................................................................................................................................. 83
4.4 RECEIVER CHAIN FOR R-OFDM ...................................................................................................................... 84
4.4.1 Multi Dimensional Demodulator ................................................................................................................ 84
4.4.2 Twin Turbo Decoder ................................................................................................................................... 84
4.4.3 Receive Chain for R-OFDM employing MIMO .......................................................................................... 84
- 6-
List of Tables
TABLE 1: OFDM SYMBOL NUMEROLOGY ________________________________________________________ 11
TABLE 2: OFDM SUPERFRAME NUMEROLOGY ___________________________________________________ 12
TABLE 3: DESCRIPTION OF FORWARD LINK CHANNELS FOR HKLLMNQRSUZ FDD __________________ 19
TABLE 4: FORWARD LINK PACKET FORMATS – SISO MODE _______________________________________ 23
TABLE 5: STRUCTURE OF DIFFERENT FL CONTROL CHANNEL MESSAGES __________________________ 24
TABLE 6: DESCRIPTION OF RL CHANNELS FOR HKLLMNQRSUZ FDD _______________________________ 36
TABLE 7: REVERSE LINK PACKET FORMATS _____________________________________________________ 37
TABLE 8: GROUP SETUP MESSAGE FIELDS __________________________________________________ 62
- 7-
List of Figures
FIGURE 1: FDD SUPERFRAME STRUCTURE 12
FIGURE 2: SUPERFRAME STRUCTURE FOR THE HALF DUPLEX MODE 13
FIGURE 3: FUNCTIONAL BLOCK DIAGRAM OF AN OFDM TRANSMITTER 14
FIGURE 4: FUNCTIONAL BLOCK DIAGRAM OF A R-OFDM TRANSMITTER 14
FIGURE 5: DRCH ASSIGNMENT 16
FIGURE 6: BRCH ASSIGNMENT 17
FIGURE 7: MULTIPLEXING MODES 18
FIGURE 8: FORWARD LINK CHANNELS FOR HKLLMNQRSUZ FDD 19
FIGURE 9: COMMON/AUXILIARY PILOT CHANNEL FOR SISO, SIMO AND MIMO SUPPORT 20
FIGURE 10: FORWARD LINK FREQUENCY HOPPING – 512PT FFT AND ABOVE 21
FIGURE 11: FORWARD LINK PILOT FORMATS: 512PT FFT AND ABOVE 22
FIGURE 12: FORWARD LINK PILOT FORMATS: 128PT AND 256PT FFT 22
FIGURE 13: EXAMPLE OF EFFECTIVE ANTENNAE TRANSMISSION.UNDER STUDY) 28
FIGURE 14: SCW MIMO – TRANSMITTER 29
FIGURE 15: EXAMPLE OF SCW MIMO – RECEIVER 30
FIGURE 16: MCW MIMO - TRANSMITTER 31
FIGURE 17: EXAMPLE OF MCW MIMO - RECEIVER 32
FIGURE 18: REVERSE LINK FREQUENCY HOPPING 34
FIGURE 19: REVERSE LINK PILOT FORMATS 35
FIGURE 20: RL CHANNELS FOR HKLLMNQRSUZ FDD 36
FIGURE 21: MULTIPLEXING CDMA AND OFDMA TRAFFIC 38
FIGURE 22: REVERSE LINK CDMA CONTROL SEGMENT 39
FIGURE 23: R-ACKCH ID ASSIGNMENTS 41
FIGURE 24: R-ACKCH STRUCTURE: 512PT FFT AND ABOVE 41
FIGURE 25: R-ACKCH STRUCTURE: 256PT FFT AND 128PT FFT 42
FIGURE 26: SUPERFRAME PREAMBLE STRUCTURE FOR FFT SIZES OF 512 AND ABOVE 45
FIGURE 27: ACCESS PROBE SEQUENCES. NS SEQUENCES WITH NP PROBES PER SEQUENCE 50
FIGURE 28: FORWARD LINK H-ARQ TIMELINE 50
FIGURE 29: FORWARD LINK H-ARQ TIMELINE FOR EXTENDED FRAME TRANSMISSION 51
FIGURE 30: ALTERNATIVE FORWARD LINK H-ARQ TIMELINE: SIX INTERLACES 51
FIGURE 31: ALTERNATE FORWARD LINK H-ARQ TIMELINE: FIVE INTERLACES 52
FIGURE 32: REVERSE LINK H-ARQ TIMELINE 52
FIGURE 33: REVERSE LINK H-ARQ TIMELINE FOR EXTENDED FRAME TRANSMISSION 53
FIGURE 34: ALTERNATIVE REVERSE LINK H-ARQ TIMELINE: SIX INTERLACES 53
FIGURE 35: ALTERNATIVE REVERSE LINK H-ARQ TIMELINE: FIVE INTERLACES 54
FIGURE 36: FORWARD LINK H-ARQ TIMELINE FOR HALF DUPLEX MODE 54
FIGURE 37: REVERSE LINK H-ARQ TIMELINE FOR HALF DUPLEX MODE 55
FIGURE 38: EXAMPLE CHANNEL TREE 57
FIGURE 39: DATA MODULATION IN THE PRESENCE OF DYNAMIC SHARING 60
FIGURE 40. EXAMPLE SET OF TIME DOMAIN RESOURCES 61
FIGURE 41: EXAMPLE OF GROUP RESOURCE ALLOCATION SIGNALING 64
FIGURE 42: REVERSE LINK TRAFFIC CHANNEL POWER CONTROL 69
FIGURE 43: SUBBAND SCHEDULING 73
- 8-
FIGURE 44: MULTIPLE CHANNEL TREES TO SUPPORT QUASI-ORTHOGONAL OPERATION 74
FIGURE 45: PILOT AND DATA SYMBOL PLACEMENT IN A TIME-FREQUENCY TRAFFIC BLOCK 75
FIGURE 46: CODING AND MODULATION STRUCTURE 81
- 9-
1 INTRODUCTION
In general the design objectives of Huawei-KDDI-LG Electronics-Lucent-Motorola-Nortel-
Qualcomm-RITT-Samsung-China Unicom-ZTE (HKLLMNQRSUZ) proposal is to provide
performance enhancements over existing cellular systems, while maintaining competitive edge over
current and emerging wireless technologies [1]. The main areas of performance improvements
include the introduction of higher peak date rates, better spectral efficiency, lower latency,
improved terminal battery life and higher capacity and enhanced user experience for delay-
sensitive applications (e.g., VoIP, video telephony). Please note that the system is still being
designed, so many of the sections are marked as “Under Study”. In these sections, the general
principle of the system is described, however the details are still being completely designed. One
sample design has sometimes been outlined, but these may change in due course.
The system design described here is a unified design for full & half duplex FDD modes of operation,
with support for scalable bandwidth. The system is designed for robust mobile broadband access,
and is optimized for high spectral efficiency and short latencies using advanced modulation, link
adaptation and multi-antenna transmission techniques. Features necessary for mobile operation
such as fast handoff, fast power control, inter-sector interference management are integrated into
the design. Adaptive coding and modulation with synchronous H-ARQ and turbo coding is used for
achieving high spectral efficiency. Subband scheduling provides enhanced performance on both the
Forward Link (FL) and the Reverse Link (RL) by exploiting multi-user diversity gains for latency
sensitive traffic.
The retransmission latency is quite small, approximately 7 ms on FL and RL. The peak rate is over
260 Mbps in a 20 MHz FL, and 70 Mbps in a 20 MHz RL.
The system employs an Orthogonal Frequency Division Multiple Access (OFDMA) FL with antenna
techniques like multiple input multiple output (MIMO) and Spatial Division Multiple Access (SDMA)
support. The RL is quasi-orthogonal. i.e. it employs orthogonal transmission based on OFDMA,
together with non-orthogonal user multiplexing with layered superposition (LS-OFDMA) or multiple
receive antennae (SDMA). The RL also includes optional CDMA transmission for low-rate traffic.
Interference management is obtained through fractional frequency reuse (FFR), improving coverage
& edge user performance. FFR is under study. Optimized throughput / fairness tradeoff is obtained
through distributed power control based on other cell interference. The RL employs a CDMA
control segment with statistical multiplexing of the various control channels and the number of
control channels reported from access terminal (AT) is adapted based on the available RL traffic to
pilot ratio. The system employs fast access with reduced overhead and fast request. The RL
employs a broadband reference signal for power control and subband scheduling. Efficient handoff
support is also provided.
Multiple antenna techniques include FL precoding & SDMA, MISO / MIMO closed loop precoding
with low-rate feedback. Single codeword (SCW) MIMO schemes with closed loop rate & rank
adaptation is supported, while multi-codeword (MCW) or layered MIMO with per-layer rate
adaptation is also a component technology.
- 10-
1.1 Forward Link of HKLLMNQRSUZ FDD
The FL of HKLLMNQRSUZ supports two forms of OFDM data transmission – traditional OFDM, and
rotational OFDM (R-OFDM). The support of R-OFDM is optional at the AT and the access network
(AN).
A basic OFDM symbol (of 113.93 s) supports five basic FFT sizes of 128, 256, 512, 1024 and 2048
corresponding to different sampling rates (see Table 1), such that the subcarrier spacing is 9.6KHz.
Flexible bandwidth operation is provided by using different numbers of guard carriers.
Parameter 128pt FFT 256pt FFT 512pt FFT 1024pt FFT 2048pt FFT Unit
Sampling rate 1.2288 2.4576 4.9152 9.8304 19.6608 Mcps
Subcarrier spacing 9.6 9.6 9.6 9.6 9.6 kHz
Bandwidth of ≤ 1.25 1.25 – 2.5 2.5 - 5 5 - 10 10 - 20 MHz
operation
Guard carriers Function of Function of Function of Function of Function of
bandwidth bandwidth bandwidth bandwidth bandwidth
Cyclic prefix 6.51-26.04 6.51-26.04 6.51-26.04 6.51-26.04 6.51-26.04 s
Windowing 3.26 3.26 3.26 3.26 3.26 s
duration
OFDM Symbol 113.93 113.93 113.93 113.93 113.93 s
duration
(for 6.51 s CP)
Table 1: OFDM Symbol Numerology
Larger subcarrier spacing is under study. Additional OFDM symbol numerologies optimized for
indoor channels is under study. OFDM symbols are transmitted within a structure consisting of a
number of frames embedded per superframe, which form the basic unit of transmission.
1.1.1 FDD Superframe
A FL superframe consists of a superframe preamble followed by a certain number of Physical Layer
(PHY) frames. The number of PHY frames in a superframe is determined by the FFT size, as per
Table 2. For an FFT size of 256 or above, an FL superframe contains 24 PHY frames (see Figure 1).
In the case of a 128 point FFT, an FL superframe contains 48 PHY Frames.
Each PHY frame consists of 8 OFDM symbols. Depending on the FFT size, each superframe
preamble carries 8, 16 or 32 OFDM symbols, as shown in Table 2. The superframe preamble
carries acquisition pilots and overhead channels for initial acquisition (see Section 1.3.1). The
OFDM symbols in the superframe preamble use a fixed cyclic prefix duration of (roughly) 26 s.
The OFDM symbols in the PHY frames use a cyclic prefix duration that is configurable from 6.5 s
to about 26 s. The details of the PHY frame duration, superframe preamble duration and
preamble duration for different FFT sizes are given in Table 2, assuming a cyclic prefix duration of
- 11-
6.51s for symbols in the PHY frames. This nominal value is used throughout this document,
although the cyclic prefix duration is a flexible parameter and can be changed.
FL PHYFrame Index 0 1 2 3 22 23 24 25 26 27
FL PHYFrame Index 0 1 2 3 22 23 24 25 26 27
AN
AN
Superframe
Superframe
Preamble
Preamble
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
PHY
PHY
PHY
PHY
Frame Frame
Frame Frame
…
…
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
Superframe
Superframe
Preamble
Preamble
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
…
…
FL Superframe
FL Superframe
RL Superframe
RL Superframe
AT
AT
PHY Frame
PHY Frame
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
…
… PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
PHY Frame
PHY Frame
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
PHY
PHY
Frame
Frame
…
…
RL PHYFrame Index 0 1 2 3 22 23 24 25 26 27
RL PHYFrame Index 0 1 2 3 22 23 24 25 26 27
time
time
Figure 1: FDD Superframe Structure
128 pt 256 pt 512 pt 1024 pt 2048 pt
Parameter Units
FFT FFT FFT FFT FFT
PHY Frame OFDM
8 8 8 8 8
Duration Symbols
(For 6.51 μs
CP) 911.46 911.46 911.46 911.46 911.46 s
Superframe OFDM
32 16 8 8 8
Preamble Symbols
Duration
(26.04 μs CP) 4.27 2.14 1.07 1.07 1.07 ms
PHY
Superframe 48 24 24 24 24
Frames
Duration
48.02 24.01 22.94 22.94 22.94 ms
Number of H-
PHY
ARQ interlaces 8 8 8 8 8
Frames
(FL & RL)
Retransmission PHY
8 8 8 8 8
Interval Frames
(FL & RL) 7.29 7.29 7.29 7.29 7.29 ms
Table 2: OFDM Superframe Numerology
The system also supports the half duplex mode as described in Section 1.3.4.3. This mode
supports AT which are not capable of receiving and transmitting at the same time. Both half-
duplex interlaces share a common superframe preamble. The superframe structure for the half
duplex mode is shown in Figure 2.
- 12-
FL PHYFrame Index 0 0' 1 1' 11 11' 12 12' 13 13'
FL
Superframe
Preamble
PHY
Frame
PHY
Frame
PHY
Frame
PHY
Frame
… PHY
Frame
PHY
Frame
Superframe
Preamble
PHY
Frame
PHY
Frame
PHY
Frame
PHY
Frame
…
FL Superframe
RL Superframe
RL
PHY
Frame
PHY
Frame
PHY
Frame
PHY
Frame
… PHY
Frame
PHY
Frame
PHY
Frame
PHY
Frame
PHY
Frame
PHY
Frame
…
RL PHYFrame Index 0 0' 1 10' 11 11' 12 12' 13
time
PHY Half-duplex PHY Half-duplex
Frame interlace 0 Frame interlace 1
Figure 2: Superframe Structure for the Half Duplex Mode
The set of PHY Frames are divided into two half-duplex interlaces. An AT operating in the half-
duplex mode receives FL PHY frames on one half-duplex interlace, and transmits RL PHY frames
on the other half-duplex interlace. There is no RL transmission corresponding to the superframe
preamble transmission on the FL. The number of PHY Frames in the superframe as well as the
duration of the superframe preamble is the same as in the full-duplex operation for a given FFT
size. Although not shown in Figure 2, additional guard time is inserted between consecutive PHY
Frames, and hence the superframe duration in half-duplex mode will be different from that in full-
duplex mode.
1.1.2 Rotational OFDM
Rotational-OFDM (R-OFDM) is an optional scheme at AT and AN, that has the potential to improve
performance in the cases where high code rates are employed over highly frequency selective
channels.
1.1.2.1 Overview of Rotational OFDM
In this scheme, the N modulation symbols that are to be transmitted over N tones of an OFDM
symbol are rotated pairwise, and mapped to a sequence of N transmission symbols. To try to
achieve independent fading across the tones for maximum diversity gain, the rotated pairwise
symbols are mapped to N tones spaced sufficiently apart in the frequency domain. For un-coded
systems, or for systems employing weak channel codes (i.e. higher code rates), the diversity gain is
significant.
At the transmitter, after the modulation symbols are code-multiplexed, the OFDM symbols are
rotated and input to the IFFT block. The cyclic prefix and windowing operations remain
unchanged. At the receiver, using an MMSE channel estimator causes interference among the
different tones of a rotation group. To minimize this effect, more sophisticated demodulation
schemes like Maximum-Likelihood-Detection (MLD) or Multi-Dimensional-Demodulator (MDD) can
be used for some additional complexity. The remaining receive chain stays unchanged.
An alternative decoding scheme, called the twin Turbo decoder, attains better performance than
MMSE/MLD demodulation followed by Turbo decoding. The twin Turbo decoder scheme is based
on iterative de-mapper, which is similar in principle to the Turbo decoder. In this scheme, the a
posteriori probabilities from the Turbo decoder output are input to the de-mapper for updating the
channel value of the transmitted bits when computing the LLRs.
- 13-
Detailed explanations on the MDD and twin Turbo decoders can be found in Section 4.4.
1.1.2.2 R-OFDM Transmitter
Figure 3 and Figure 4 represent the functional block diagrams of a normal OFDM and a R-OFDM
transmitter respectively.
x1
x2 x1 F1
ENC MOD S/P IFFT
(Turbo) F2
x2
Figure 3: Functional Block Diagram of an OFDM Transmitter
x1 y1
x2 x1 F1
ENC MOD S/P RCM IFFT
(Turbo) F2
x2 y2
Figure 4: Functional Block Diagram of a R-OFDM Transmitter
The difference between the two transmitters is the Rotational-Code-Multiplexer (RCM), which
generates the rotated symbols using the following equation:
y1 r11 r21 rD1 x1
y r r22 rD 2 x2
2 12 , (i.e., y = Rx ),
y D r1D r2 D rDD x D
where the vectors x and y represent the modulated symbols and the rotated symbols respectively.
The matrix RD represents the rotational codes. Note that D denotes “rotational dimension.” The
rotational dimension indicates the number of modulation symbols mapped to different tones. For a
rotational dimension of D, the rotational matrix RD is given by:
R cos D R D / 2 sin D
RD D/2 . Sample instantiations of R for D=2 and D=4 are given by:
R D / 2 sin D R D / 2 cos D
r r21 cos 2 sin 2
R 2 11 and
r12 r22 sin 2
cos 2
- 14-
cos 2 cos 4 sin 2 cos 4 cos 2 sin 4 sin 2 sin 4
R cos 4 R 2 sin 4 sin 2 cos 4 cos 2 cos 4 sin 2 sin 4 cos 2 sin 4
R4 2 .
R 2 sin 4 R 2 cos 4 cos 2 sin 4
sin 2 sin 4 cos 2 cos 4 sin 2 cos 4
sin 2 sin 4 cos 2 sin 4 sin 2 cos 4 cos 2 cos 4
The rotated symbols obtained by the above process are then input to the serial-to-parallel converter
(S/P) followed by the IFFT, at which point the transmit chain is identical to a conventional OFDM
transmitters.
1.1.3 Forward Link Resource Channel
The proposed system supports two resource channels on the FL data channel, viz. Distributed
Resource Channel (DRCH) and Block Resource Channel (BRCH) to efficiently support both
frequency diversity and advanced transmission techniques such as precoding interference
estimation and FL power control.
1.1.3.1 DRCH
DRCH consists of a set of tones scattered across entire bandwidth available to the DRCH. DRCH (N,
k) consists of regularly spaced T/N sub-carriers and 8 contiguous OFDM symbols, where T denotes
the total number of useful sub-carriers in each OFDM symbol. T/N is 16 for data channel. The
T/N sub-carriers in each OFDM symbol is referred to as hop port. Hop permutation maps the hop
port to frequency and changes every OFDM symbol. The hop permutation can be independent
across sectors. An example of DRCH assignments shown in Figure 5.
- 15-
A PHY frame
(8 OFDM symbols)
User 1 with DRCH(30, 0)
User 2 with DRCH(30, 3)
sub-carrier
N = 30
freq
. .
time
. .
. .
(a) Example of DRCH (30, k) (b) Example of DRCH assignment
. for STBC support
Figure 5: DRCH Assignment
1.1.3.2 BRCH
In a system with 512pt FFT and above, users are assigned sets of 16 contiguous subcarriers that
are distributed randomly across frequency as shown in Figure 4. The mapping between hop ports
and frequency is kept constant through out the physical layer frame (8 OFDM symbols). Each set
therefore defines a hop region consisting of 16 contiguous subcarriers and 8 contiguous OFDM
symbols, also referred to as a tile. The hop permutation changes every PHYFrame and is
independent across sectors.
In a system with 128pt or 256pt FFT, minimum assignment of 16 subcarries over 8 OFDM symbols
is split in two contiguous sets of 8 subcarriers over 8 OFDM symbols as shown in Error!
Reference source not found.Error! Reference source not found.. Hence, minimum assignment
consists of two disjoint hop regions that hop independently. The goal of this design is to provide an
adequate diversity for minimum assignment which allows us to allocate FL signaling resources
with a smaller minimum size and granularity thereby reducing control overhead.
- 16-
A PHY frame
(8 OFDM symbols)
User 1
User 2
User 3 8 sub-carriers
sub-carrier
16 sub-carriers
. .
. .
. .
freq
time
(a) Example of BRCH assignment (b) Example of BRCH assignment
for 512pt FFT and above for 128pt and 256pt FFT
Figure 6: BRCH Assignment
1.1.4 Multiplexing of Resource Channel
DRCH and BRCH assignments coexist in the same PHY Frame so that frequency diversity and
frequency selective transmissions are simultaneously supported. There are two multiplexing modes,
Mode 1 and Mode 2.
1.1.4.1 Multiplexing Mode 1
In this mode, DRCH punctures BRCH (see Error! Reference source not found.Error! Reference
source not found.). Overhead channel indicates how many DRCHs are used so that ATs know the
puncturing pattern. In this mode, common pilot is transmitted over the entire bandwidth.
Additional dedicated pilot is TBD. Support by the AT for an assignment that includes both DRCH
elements and BRCH elements to the same AT. is TBD.
1.1.4.2 Multiplexing Mode 2
In this mode, DRCH and BRCH are only used on different sub-bands. The overhead channel
indicates how many sub-bands are in DRCH and how many sub-bands are in BRCH. The size of a
subband is configurable by the BS through the overhead channel. Support for an assignment or a
- 17-
set of simultaneous assignments that include both DRCH elements and BRCH elements to the
same AT is optional at the AT.
A PHY frame
(8 OFDM symbols)
DRCH assignments
BRCH assignments
BRCH Subband
. .
. .
. .
DRCH Subband
freq
time
(a) Multiplexing Mode 1 (b) Multiplexing Mode 2
Figure 7: Multiplexing Modes
1.1.5 Forward Link Channels
The FL channels are shown in Figure 8, and described in Table 3.
F-PBCCH F-SBCCH F-DCH F-CPICH
F-ACQCH F-OSICH F-DPICH F-AuxPICH
PHY
Channels
F-SSCH F-PCCH F-FOSICH F-PQICH
- 18-
Figure 8: Forward Link Channels for HKLLMNQRSUZ FDD
F-PBCCH Primary Broadcast Channel, Carries Deployment Specific Parameters
F-SBCCH Secondary Broadcast Channel, Carries Sector Specific Parameters
F-OSICH Broadcast Channel, Carries Indication of Inter-Sector Interference
F-ACQCH Acquisition Channel
F-CPICH Common Pilot Channel
F-DPICH Dedicated Pilot Channel
F-AuxPICH Auxiliary Common Pilot Channel
F-SCCH Shared Signaling Channel, Carries Access Grants, Assignment Messages, and
Other Messages Related to Resource Management
F-PCCH Power Control Channel, carries RL power control commands per AT
F-ACKCH Acknowledgement Channel: acknowledgement bits for RL H-ARQ
F-PQICH Pilot Quality Indicator carries strength of RL pilot (R-PICH) per AT
F-FOSICH Broadcast channel indicating inter-sector interference level, complementary
wrt. F-OSICH, allows for a faster loading control with a limited penetration
F-DCH Data (Traffic) Channel
Table 3: Description of Forward Link Channels for HKLLMNQRSUZ FDD
The congestion control channel for CDMA is under study.
1.1.5.1 Forward Link Pilot Channel
Forward pilot channels are used at least for CQI estimation and data demodulation. Forward pilot
channels consist of forward common pilot channel (F-CPICH) and forward dedicated pilot channel
(F-DPICH). Pilot symbols puncture the resources in DRCH and BRCH. The usage of F-CPICH and
F-DPICH depends on the multiplexing modes. In multiplexing mode 1, F-CPICH is used for both
CQI estimation and data demodulation. Additional F-DPICH for data demodulation is under study.
In multiplexing mode 2, F-CPICH is used for both CQI estimation and data demodulation in DRCH
subbands, and F-DPICH is used for data demodulation in BRCH subbands. An additional
broadband F-CPICH may be used in BRCH subbands to enable CQI estimation and support
precoding.
1.1.5.1.1 Forward Common Pilot Channel (F-CPICH)
Common pilot placed on equi-spaced tones both in frequency and time domain. Pilot staggering is
used. The staggering provides excess delay spread mitigation. The pilot channel power is set so as
to provide sufficient channel estimation resources to a cell edge user – it has flexible power and
bandwidth overhead.
- 19-
1.1.5.1.1.1 MIMO Support with F-CPICH
An auxiliary pilot is used for MIMO support. Common and auxiliary pilot placed on equi-spaced
tones both in frequency and time domain. Pilot staggering over 4 OFDM symbols is used. Common
and auxiliary pilots are time and frequency multiplexed. The common pilot and the auxiliary pilot
have the same density of 3.125%. The pilot position in frequency domain is offset from one sector
to another sector to allow for pilot boosting which can improve the channel estimation
performance. The density of auxiliary pilots can be further optimized (TBD).
A PHY frame
(8 OFDM symbols)
pilot for ant 1 (Common Pilot)
pilot for ant 2 (Auxiliary Pilot)
8 sub-
carriers
pilot for ant 3 (Auxiliary Pilot)
pilot for ant 4 (Auxiliary Pilot)
.
.
.
Figure 9: Common/Auxiliary Pilot Channel for SISO, SIMO and MIMO Support
1.1.5.1.2 Forward Dedicated Pilot Channel (F-DPICH)
The pilot symbols populate a pre-defined pattern on the time-frequency grid. Three pilot patterns
are depicted in Error! Reference source not found.Error! Reference source not found.. Format
0 is the default pilot pattern and is used while transmitting to users capable of supporting up to
three spatial streams1. Format 1 is used to support users with high delay spread channels and
Format 2 is used to support four spatial streams. The pilot symbols are used for interference
estimation as well as to estimate the channel over every data symbol within a hop region using
time and frequency interpolation. Hop permutation maps assigned hop ports to the pilot tiles in
frequency, and is fixed for duration of frame. The hopping is independent across sectors.
1 Multiple spatial streams result from MIMO and/or SDMA transmission.
- 20-
Channel and interference estimation must be performed over every tile, as the channel can be
thought to be localized – thus a dedicated pilot is needed. The interference level is roughly constant
across a tile in a synchronized system. The different pilot patterns provide enough “looks” to
capture time & frequency selectivity by trading-off pilot overhead with support for MIMO and high
delay spread channels. The pilot pattern indicated through the packet format, and the pilots and
data symbols within block undergo the same transmit processing. The pilot overhead is identical
for SIMO and MIMO, and is 14.06% for Format 0 and 18.75% for formats 1 and 2. Further
optimization of the pilot patterns is under study.
16 tones
8 OFDM symbols
symbol
……..
OFDM
User 2 User 1 User 3 User 2
……..
……..
User 1 User 2 User 2 User 3
……..
16 sub-carriers
Frequency
Data symbol F-DPICH (Format 0) F-DPICH (Format 1) F-DPICH (Format 2)
Figure 10: Forward Link Frequency Hopping – 512pt FFT and above
1.1.5.1.2.1 MIMO Support with F-DPICH
The dedicated pilot channel consists of orthogonal overlapped pilot sequences over each contiguous
pilot cluster – as an example, consider the six strips of pilots placed as depicted in Error!
Reference source not found.Error! Reference source not found.. Each strip consists of 3 or 4
contiguous pilots. The former can support up to third order of spatial multiplexing while the latter
can support up to fourth order spatial multiplexing. The orthogonal sequences associated with
different effective antennae are defined by the columns of the DFT matrix of size 3, 2 and 4 for
Format 0, 1 and 2 respectively.
- 21-
8 OFDM symbols 8 OFDM symbols 8 OFDM symbols
cluster
cluster
16 tones
16 tones
16 tones
Format 0 18 pilots Format 1 24 pilots Format 2 24 pilots
SIMO + MIMO (rank = 3) High delay spread MIMO (rank = 4)
Data symbol Pilot symbol
Figure 11: Forward Link Pilot Formats: 512pt FFT and above
There are two pilot patterns applicable in this case are specified by Format 0 and Format 1 in
Figure 11Error! Reference source not found.Error! Reference source not found.. Format 0
defines the default pilot pattern which can be used to support spatial multiplexing orders 1
through 3. Format 1 is used to support users with high delay spread channels and can support
spatial multiplexing orders 1 and 2.
8 OFDM symbols 8 OFDM symbols
8 tones
8 tones
cluster
Format 0 18 pilots Format 1 24 pilots
SIMO + MIMO (rank = 3) High delay spread
Data symbol Pilot symbol
Figure 12: Forward Link Pilot Formats: 128pt and 256pt FFT
1.1.5.2 F-DCH
The traffic channel carrying data is called Forward Data Channel (F-DCH). F-DCH uses a rate 1/3
convolutional code for block lengths less than 128 and a rate 1/5 turbo code for block lengths
greater than 128. The code is punctured or repeated to achieve desired code rate. It uses
synchronous H-ARQ. The channel interleaver is based on a bit-reversal pattern providing almost-
regular puncture patterns and good interleaver distance properties at all code rates.
- 22-
F-DCH supports QPSK, 8PSK, 16QAM, and 64QAM - this allows a wide range of spectral
efficiencies. To avoid repetition of coded bits, the F-DCH uses modulation step-down at high
spectral efficiencies, and gains up to 1 dB for later transmissions. Table 4 lists the packet formats
for F-DCH for SISO transmissions.
Spectral Spectral Max Modulation order for each
Packet
efficiency on efficiency on number of transmission
Format
1st trans- 2nd trans- trans-
Index 1 2 3 4 5 6
mission mission missions
0 0.2 6 2 2 2 2 2 2
1 0.5 6 2 2 2 2 2 2
2 1.0 6 2 2 2 2 2 2
3 1.5 6 3 2 2 2 2 2
4 2.0 6 4 3 3 3 3 3
5 2.5 6 6 4 4 4 4 4
6 3.0 6 6 4 4 4 4 4
7 4.0 6 6 6 4 4 4 4
8 5.0 6 6 6 4 4 4 4
9 6.0 3.0 6 6 6 4 4 4 4
10 Non-decodable 3.5 6 6 6 4 4 4 4
11 Non-decodable 4.0 6 6 6 6 4 4 4
12 Non-decodable 4.5 6 6 6 6 4 4 4
13 Non-decodable 5.0 6 6 6 6 6 4 4
14 Non-decodable 5.5 6 6 6 6 6 4 4
15 NULL NULL
Table 4: Forward Link Packet Formats – SISO Mode
Note that packet formats 10 - 15 are non-decodable at the first transmission, but allows for high
spectral efficiency at target termination. Further details on the transmit chain of this channel can
be found in Section 4.2.
1.1.5.3 Forward Link Control Channels
FL control channels are transmitted over an FDM control segment in every FL PHY Frame and
include F-SCCH which is used to assign and manage FL and RL resources and specify the
respective packet formats and grant access to users in the idle state, F-ACKCH which is used to
acknowledge RL transmissions, F-PCCH which is used to send RL power control commands, F-
PQICH which is used to send RL pilot quality indicators, and F-FOSICH which is used to broadcast
other sector interference indications. The need for RL Congestion Control channel and grouped
signaling and encoding for efficient resource management (especially VoIP) are under study (TBD).
- 23-
In systems with 512pt FFT and above, FL control segment is allocated a minimum of three DRCH
or BRCH base nodes with possible increments of one base node on the channel tree via signaling
on the FL primary broadcast channel. Thus, in a 5MHz deployment, this amounts to about 10%
minimum bandwidth overhead with granularity of 3.3%. In systems with 128pt and 256pt FFT,
minimum FL control segment allocation is two base nodes with possible increments of one base
node. Modulation symbols of each channel within the entire FL control segment are interleaved
across all base nodes assigned to this segment to ensure maximum diversity. With the present
design, frequency diversity of the entire bandwidth is captured when the FL control segments are
transmitted with DRCH. When the control segments are transmitted with BRCH, at least third
order diversity is achieved with the minimum FL control segment allocation in systems with 512 pt
FFT and above, and at least fourth order diversity in systems with 128 pt and 256 pt FFT. FL
control allocation parameters such as the total bandwidth allocation, sizing of the five channel (F-
SCCH, F-ACKCH, F-PCCH, F-PQICH, and F-FOSICH), number of modulation symbols per control
channel message and a suitable packet format, are periodically broadcast along with other sector
parameters. Furthermore, periodicity of AT specific channels, namely F-PCCH and F-PQICH is
adjusted per AT via upper layer signaling. The unicast signaling enables overhead minimization
particularly when targeting users with widely varying channel quality, as the power overhead is
flexible, and adjustable every PHY Frame. This allows the system to tailor the overhead required for
signaling for a variety of usage scenarios, and no power overhead is wasted when signaling needs
change rapidly. In the following paragraphs, we describe the purpose and structure of various
channels of the FL control segment.
1.1.5.3.1 Forward Link Shared Control Channel (F-SCCH)
The FL Shared Control Channel (F-SCCH) carries a number of control channel messages that
allocate or de-allocate different resources to or from a given AT. A selection of FL control channel
messages is shown in Table 5. Columns of this table indicate different fields while rows correspond
to different control channel messages. Every cell in the table indicates multiplicity of a given field.
A 3-bit block type field allows the AT to identify the type of message and therefore interpret the
subsequent fields. The set of information bits of every message is extended by a 16-bit CRC to
enable reliable detection.
- 24-
Block Persis- Ext.
Field MACID ChanID PF Duration Timing Suppl. Rank
type tent TX
# bits 3 9-11 1 6-8 4-6 2 1 6 1 2
Access Grant 000 1 0 0 0 0 0 1 0 0
FLAM 010 0 1 1 1 0 1 0 1 0
RLAM 011 0 1 1 1 0 1 0 1 0
TBD TBD TB TBD TBD TBD TBD TBD
MCW FLAM1 100 TBD
D
TBD TBD TBD TB TBD TBD TBD TBD TBD
MCW FLAM2 101
D
TBD TBD TBD TB TBD TBD TBD TBD TBD
SCW FLAB 110
D
Table 5: Structure of Different FL Control Channel Messages
The first message carries the Access Grant message and is used to acknowledge an access attempt
by an AT, assign a new MACID, and supply a 6-bit timing adjustment for the AT to align its RL
transmission with the RL timing of the AN. The sequence of modulation symbols corresponding to
the Access Grant is scrambled according to the index of the preceding access probe transmitted by
the AT, enabling the AT to respond only to Access Grant messages that correspond to the probe
sequence that it transmitted. The FL and RL assignment messages all have a “persistent” bit that
indicates whether the assignment is only for one packet, or lasts until explicitly de-assigned or lost
due to packet failure(s). All of these assignment messages are also scrambled with the MACID of
the target AT.
The FL assignment message contains a channel ID (ChanID) that indicates the hop-ports assigned
(via the channel tree in use), the packet format (PF) to use (specifies modulation, coding, and
dedicated pilot format), the number of PHY frames occupied by this assignment, and an indication
(denoted by Ext TX) of whether or not to use the extended transmission duration for the
assignment (in extended transmission, each H-ARQ transmission spans multiple PHYFrames).
The FL assignment message (FLAM) signals a FL resource assignment to an active AT with
resources assigned indicated by ChanID and spectral efficiency indicated by PF. The new field in
this message is the supplemental assignment flag. Whenever set, this flag indicates an incremental
assignment if ChanID is not part of the existing assignment and decremental assignment
otherwise. Such a supplemental assignment takes effect starting from the new packet. Note that
the new assignment replaces the existing one if supplemental flag is not set. The RL assignment
message (RLAM) signals RL resource assignments in a fashion identical to FLAM.. Definitions of
MCW/SCW FLAM is under study (TBD).
The multi-codeword FLAM is a FL assignment message that can be used for ATs in the MIMO
multi-codeword mode. The single-codeword MIMO FLAM is similar to the FLAM.
Grouped/multicast signaling and encoding for efficient resource assignment is under study.
- 25-
1.1.5.3.1.1 Summary of Forward Link Control Channel Messages
It is easy to see that most messages span between 31 and 34 bits including 16-bit CRC when
ChanID spans 6 bits (between 33 and 36 bits when ChanID spans 8 bits). Based on this
observation, all messages can be padded to the same maximum number of bits (34 or 36,
depending on ChanID size) with a relatively low efficiency loss. Having a unified size for all
signaling messages is convenient when all the messages are encoded and modulated separately
since it removes the need to use extra overhead to indicate message sizes.
1.1.5.3.1.2 Modulation of Forward Link Control Channel Messages
All control channel messages will be independently encoded and modulated with the same spectral
efficiency, hence resulting in the same number of modulation symbols. Rate 1/3 convolutional
encoder with constrained length 9, appropriate puncturing and QPSK modulation will be used to
achieve the desired spectral efficiency. The number of modulation symbols per message (hence
spectral efficiency) is a quasi-static parameter specified through an overhead message. Spectral
efficiencies on the order of 0.5-1 bps/Hz will be used. Furthermore, every message will be power
controlled individually according to the FL channel strength of the target ATs. Such a design allows
for a low bandwidth overhead with flexible power overhead that depends on the instantaneous
F-SCCH load and can be adjusted within every FL PHY Frame.
1.1.5.3.2 Acknowledgement Channel (F-ACKCH)
This channel is used to acknowledge RL H-ARQ transmissions and therefore is present in every FL
PHY frame to acknowledge the associated RL PHY frame. Each acknowledgement (ACK) is a
message, indicating either a NACK or ACK. The maximum number of ACK bits required is equal to
the total number of usable channel IDs since ACKs are linked to a channel assignment. Every base
node of the channel tree is associated with an ACK bit. A larger channel assignment corresponding
to multiple base nodes uses the ACK bit associated with the base node with the lowest channel ID
in the assignment.
ACK bits are encoded using on/off keying so that the absence of ACK transmission implies negative
acknowledgement. The ACKs are encoded over a number of modulation symbols that are placed at
different frequencies to provide channel and interference diversity. The absence of ACK
transmission implies negative acknowledgement, „+1‟ implies positive acknowledgment.
Additional ACK resources are provided for users operating on CDMA traffic segments. ACKs are
encoded over a number of modulation symbols that are placed at different frequencies to provide
channel and interference diversity. Pre-coded ACK transmission is under study (TBD).
1.1.5.3.3 Reverse Link Power Control Channel (F-PCCH)
The RL power control (PC) channel carries commands for closed loop control of RL control channel
transmit power. BPSK modulation is used to transmit power control bits.
Power control frequency around 150Hz has been found to be sufficient for most channel
conditions. This implies that every active AT could receive a power control bit once every six PHY
frames. Hence, RL power control channel of every PHY frame reports power control bits for a
subset of active terminals. Faster control rate is under study (TBD).
- 26-
1.1.5.3.4 Reverse Link Pilot Quality Indicator Channel (F-PQICH)
The RL pilot quality indicator (PQI) channel carries quantized values indicating the quality of pilots
transmitted on the RL pilot channel (R-PICH) of each active terminal. The AT uses pilot quality
indicator values from different active set members for RL serving sector selection. It also uses
these values to set the transmit powers on different RL control channels and RL traffic channel.(
1.1.5.3.5 Fast OSI Channel (F-FOSICH)
This channel is used to provide other sector interference indication to an AT from sectors in its
active set. As compared to the F-OSICH physical layer channel, this channel is transmitted at a
faster rate but with less coverage. When present, this channel carries a three-state OSI value and
occupies 8 modulation symbols in each PHY Frame. This channel therefore has a very small
overhead. The OSI value is used as part of the RL PC algorithm, which is described in 1.3.9.3. The
power level of this channel can be configured to achieve the desired penetration depth.
1.1.5.3.6 Other Channels
TBD – two control channels under consideration are the congestion control channel for the CDMA
RL, and the control signaling channel for group resource allocation on the FL (see Section 1.3.7.4).
1.1.6 MIMO Design
HKLLMNQRSUZ FDD supports multiple-input multiple-output (MIMO) techniques that increase
spectral efficiency through spatial multiplexing. This section describes the MIMO design in detail
taking into consideration factors like H-ARQ, rate prediction, channel estimation, feedback
overhead, spatial correlation effects, complexity and effect on mobility. The design simultaneously
supports both SISO, SIMO and MIMO users.
1.1.6.1 Data Channel Structure
The system uses the concept of effective antenna signaling at the base-station, i.e., the AN creates
multiple linear combinations of physical antennae, each linear combination being referred to as an
effective antenna. The different linear combinations are generated to preserve the channel statistics
as well as to transmit the same power from all physical antennae. For the first constraint, unitary
matrices are needed to generate the spatial signatures. For the second constraint, the sum of the
power of the entries for every row is constant. The exact structure of effective antennae signaling
(i.e. the structure of linear combinations) is implementation specific and is transparent to the AT.
Classic examples of effective antenna signaling include cyclic delay diversity and phase sweep
diversity however other forms can be used as well.
1.1.6.2 Pilot Structure
In this section the pilot structure for DRCH and BRCH is described.
1.1.6.2.1 DRCH
In DRCH, a common broadband pilot is transmitted from each effective antenna. Pilot subcarriers
are present in every OFDM symbol, and the set of pilot sub-carriers in each OFDM symbol are
spaced equally over the entire bandwidth to enable efficient channel estimation. A common pilot
- 27-
channel (F-CPICH) is transmitted from the first (SIMO) effective antenna and is used for SIMO
demodulation. An Auxiliary Pilot Channel (F-AuxPICH) is transmitted from each of the remaining
effective antennae. F-CPICH and F-AuxPICH subcarriers from the different effective antennae are
TDM‟d with each other. See Section 1.1.5.1.1 for the pilot structure.
1.1.6.2.2 BRCH
In BRCH mode, the assignment for a MIMO transmission consists of one or more tiles. Each tile is
16 contiguous subcarriers by 8 OFDM symbols. The pilot patterns, used to estimate the MIMO
channel, are described in Section 1.1.5.1. These patterns allow multiplexing pilots that correspond
to different spatial streams, including MIMO and/or SDMA streams. .
1.1.6.3 STTD Mode
In space time transmit diversity (STTD) mode the AN is assumed to employ up to four effective
antennae. The STTD block code is transmitted at the same subcarrier and two consecutive OFDM
symbols. The support of STTD requires the same hopping pattern across two consecutive OFDM
symbols. The need for STTD mode is under study.
1.1.6.4 MIMO Modes
Let the number of transmit antennae be M t , the number of receive antennae be M r , the number
of effective antennae be M e , and the number of modulation symbols simultaneously transmitted
on a given subcarrier and OFDM symbol (a.k.a. spatial multiplexing order) be
M min Me , Mr 2. The design supports two main MIMO modes, namely single codeword (SCW)
and multiple codeword (MCW) designs.
In the SCW mode, one codeword is transmitted in the frequency-space domain. A simple linear
receiver can be used to decouple the multiple transmitted modulation symbols although it may not
achieve capacity. More sophisticated receiver for SCW can significantly improve performance,
especially at high SNR, on the expense of increasing complexity at AT.
In the MCW mode, multiple encoded streams of data are simultaneously transmitted. A successive
interference cancellation (SIC) receiver can be used, which can achieve capacity.. However, the SIC
process can increase AT complexity as well as memory requirements. In addition, this mode
requires extra overhead compared to SCW mode as will be described later.
There are many ways of utilizing spatial and frequency diversity in a MIMO channel. As an
example, when transmitting M spatial layers over M e effective antennae for DRCH,, the mapping
between the spatial layers and the effective antennas can be cyclically shifted for different
subcarriers, as shown in Figure 13. That is, on the i-th subcarrier of the assignment, the
modulation symbol associated with the j-th spatial layer is transmitted on the k-th effective
antenna, where k (i j ) mod M e . This cyclic shift is under study – TBD.
2 The constraint of M M r is necessary if linear receivers are used to decouple the incoming
MIMO sub-streams.
- 28-
Figure 13: Example of Effective Antennae Transmission.Under Study)
Note that SISO is transmitted form the first effective antenna, while MIMO is transmitted from all
Me effective antennae. Transmitting over the physical antennae is a special case of this scheme.
1.1.6.4.1 Single Codeword (SCW) Design
In this section the SCW MIMO scheme is described in detail.
1.1.6.4.1.1 Rate and Rank Prediction
The code rate, the constellation size, and the spatial multiplexing order (also denoted by rank) M
are adapted to the channel. The AT feeds back the rank value together with the corresponding
channel quality indicator (CQI) value to AN. Specific details of the feedback are under study. The
AN adjusts the transmitted power level, based on the power control loop and rank, and runs a rate
prediction algorithm by which it chooses the packet format (PF). The PF, together with the rank,
define the data rate transmitted.
1.1.6.4.1.2 Transmitter Structure
The transmitter structure is shown in Figure 14. The input data stream is turbo encoded using the
selected code rate, and mapped to the selected QAM constellation. The stream of modulation
symbols is then de-multiplexed to M parallel sub-streams. The M sub-streams are mapped to the
physical antennae using the effective antenna signaling described in Section 1.1.6.1, thus adapting
the rate and rank to channel realizations. The H-ARQ structure is similar to SISO.
- 29-
Figure 14: SCW MIMO – Transmitter
1.1.6.4.1.3 Example of Receiver Structure
The example of receiver presented in this contribution runs a linear MMSE filter on the received
samples to decouple the incoming M sub-streams3. The soft estimates of the modulation symbols
are then fed to an LLR computer and the output is fed to a Turbo decoder. The receiver structure is
shown in Figure 15.
3 As mentioned before, more sophisticated receiver structures can be used.
- 30-
OFDM
1 Demod 1
Parallel to Serial
OFDM 2 Decoded
(per tone)
2 Demod
MMSE
Turbo bits
LLR
Decoder
M
OFDM
MR
Demod
Post-processing SNRs
CQI Rank Prediction
Rank CQI-Quantization
Figure 15: Example of SCW MIMO – Receiver
The receiver can employ more sophisticated detectors – the complexity is low if only linear MMSE is
employed.
1.1.6.4.1.4 Reverse Link Feedback Channels for SCW
The CQI and rank are reported by AT via R-MCQICH that is periodically transmitted. The exact
format of this control channel is under study (TBD). MIMO SCW transmissions are acknowledged
via R-ACKCH in the same way as non-MIMO transmissions.
1.1.6.4.2 Multi Codeword (MCW) Design
The MCW schemes are capacity achieving. For this design, the CQI is fed back for each layer, i.e.,
rate prediction is done per layer. A successive interference cancellation (SIC) receiver is used to
decouple the M layers providing higher throughput and more tolerance to spatial correlation.
Within the maximum number of H-ARQ transmissions, no new packets are transmitted on the
decoded layers. Total power is equally divided on the outstanding layers. In the MCW scheme M
codewords (or packets) are transmitted in parallel. Each codeword is transmitted form all effective
antennae to exploit the available spatial diversity.
1.1.6.4.2.1 Rate and Rank Prediction
The code rate and the constellation size on each of the M data sub-streams are adapted to
channel. The AT runs a rank prediction algorithm by which it determines the value of M to be
used. The AT also computes M CQI values, one for each data sub-stream, and feeds them back to
the AN. Specific details of the feedback are under study. The AN adjusts the transmitted power
- 31-
level on each data sub-stream, based on the power control loop and rank, and runs a rate
prediction algorithm by which it chooses the packet format for each data sub-stream.
1.1.6.4.2.2 Transmitter Structure
The transmitter structure is shown in Figure 16.M data packets are transmitted in parallel. The
m , m 1, , M , data packet is turbo encoded using the m th selected code rate, and mapped to
th
th th
the m selected QAM constellation. The modulations symbols corresponding to the m data
packet is denoted hereafter by a layer. The M layers are then mapped to the physical antennae
using the effective antenna signaling described in Section 1.1.6.1.
1 OFDM
1
Effective Antenna Signaling
st Mod
1 Data Packet Turbo QAM 1
Encoder Map
2 OFDM
2 2
2nd Data Packet Turbo QAM Mod
Encoder Map
Mth Data Packet Turbo QAM M
Encoder Map Mt OFDM Mt
Mod
CQI1
Rate AP
Prediction CQIMe Receiver
Figure 16: MCW MIMO - Transmitter
1.1.6.4.2.3 Example of Receiver Structure
The example of receiver presented in this contribution is a SIC receiver with linear MMSE filter.
The receiver attempts first to decode the first layer. The linear MMSE filter generates the soft
estimate of the modulation symbols corresponding to the first layer and all subcarriers in the
user‟s assignment. The different soft estimates are sent to an LLR computer, and the resultant
LLRs are fed to the Turbo decoder. If the first layer is decoded properly (passes the CRC), the
receiver regenerates a clean version of the modulation symbols corresponding to the first layer,
multiplies each modulation symbol by the corresponding channel coefficient, and subtracts the
contribution of the first layer from the received signal. The receiver then attempts to decode the
second layer, if decoded, the receiver subtracts its contribution from the received signal, and so on.
If at any point, one of the layers is not decoded, the receiver stops the decoding process and sends
an R-ACKCH indicating the layers that got decoded. The receiver structure is shown in Figure 17.
- 32-
OFDM
1 Demod
Layer Cancelation
OFDM
(per tone)
2
Demod
MMSE
Decoded bits
Turbo
LLR
Decoder
MR OFDM
Demod
Post-processing SNRs
CQI 1
CQI-Quantization
CQI Me
Figure 17: Example of MCW MIMO - Receiver
1.1.6.4.2.4 BL H-ARQ
q , on the q 1 transmission the AN
th
If N dec layers have been decoded at some transmission
does not transmit any new codewords on the successfully decoded layers and only sends
redundancy information on the M N dec layers that have not yet been decoded. In doing so, the
AN equally divides the available power per subcarrier on the outstanding layers. The remaining
layers are blanked.
1.1.6.4.2.5 Reverse Link Feedback Channels for MCW
The CQI information is sent on the R-MCQICH. The AT feeds back a CQI value for each layer. The
exact format of R-MCQICH is under study (TBD). Each layer of MIMO MCW transmission is
acknowledged independently. Hence, M t instances of the R-ACKCH are needed to acknowledge
M t layers. Acknowledgement of MIMO MCW assignments is explained in section 1.2.2.4.1.
1.1.6.5 Precoding
In a FDD system, the FL and RL transmissions are on widely separated frequencies. As a result,
the FL channel and the RL channel may fade independently. A direct consequence is that the RL
channel estimates do not provide instantaneous channel knowledge of the FL. Nevertheless,
transmit beamforming gains, in an FDD system are possible by explicitly feeding back FL channel
information over the RL and then using this information to transmit data in the preferred direction
to every user. We refer to this technique as precoding (a.k.a closed loop MIMO).
- 33-
1.1.6.6 Space Division Multiple Access
Space-division multiple access (SDMA) on the FL is an advanced transmission technique where
multiple users are signaled on the same time-frequency resources. A key characteristic of SDMA is
the opening up of new dimensions at the expense of reduced signal to interference and noise ratio
(SINR). Users are overlapped using beams pre-defined in codebook, providing increased system
dimension and adaptive sectorization gain. The details of SDMA schemes are under study with
precoder and feedback design.
1.2 Reverse Link of HKLLMNQRSUZ FDD
On the RL, OFDMA and CDMA traffic is frequency multiplexed, with the split between CDMA and
OFDMA being configurable by the AN, and is sector and AT specific. The use of DFT-SOFDM for RL
is under study. Both CDMA and OFDMA traffic are power controlled. The RL CDMA Traffic Channel
is used for the transmission of low-rate bursty delay-sensitive services (e.g. VoIP, Gaming etc), and
supports a limited set of transmission formats. Frequency-domain interference cancellation is used
on this channel, which uses fast power control, H-ARQ and slow distributed scheduling.
The RL OFDMA Traffic Channel is fully scheduled, and supports quasi-orthogonal multiple
antenna operation (QORL) and layered superposed OFDMA (LS-OFDMA).
Support for 2xN MIMO on the reverse link is under study.
A RL superframe consists of 24 Physical Layer (PHY) frames for FFT sizes of 256 above, see Figure
1, and of 48 RL PHY Frames for the case of a 128 point FFT. Each RL PHY frame, except for the
first RL PHY Frame in a superframe, consists of 8 OFDM symbols, and has the same duration as a
FL PHY Frame. The first RL PHY frame in a superframe is elongated so as to align FL and RL
superframes. The duration of a RL superframe is therefore the same as the duration of a FL
superframe and is given in Table 2.
The HKLLMNQRSUZ system also supports the half duplex mode as described in Section 1.3.4.3.
This mode supports ATs which are not capable of receiving and transmitting at the same time.
Both half-duplex interlaces share a superframe preamble. The superframe structure for the half
duplex mode is as shown in Figure 2.
1.2.1 Reverse Link Resource Channel
Every user is assigned sets of 16 contiguous subcarriers that are distributed randomly across
frequency. The mapping between hop ports and frequency is kept constant through out the PHY
Frame (8 OFDM symbols). Each set therefore defines a hop region consisting of 16 contiguous
subcarriers and 8 contiguous OFDM symbols. The hop permutation changes every PHY frame and
is independent across sectors.
- 34-
OFDM symbol
User 2
User 1
User 3
User 2
Frequency
User 1
User 2
16 sub-carriers
Data symbol
R-DPICH (Format 0)
R-DPICH (Format 1)
8 OFDM symbols
Figure 18: Reverse Link Frequency Hopping
Pilot patterns are chosen to provide enough “looks” at the channel to capture time & frequency
selectivity. For quasi-orthogonal RL (QORL) and LS-OFDMA support, orthogonal overlapped pilot
sequences are provided over each contiguous pilot cluster.
1.2.1.1 Dedicated Pilot Channel
The pilot symbols populate a pre-defined pattern within this two-dimensional grid. Two pilot
patterns are defined and are as depicted in Figure 18. Format 0 is the default pilot pattern. Format
1 is used by ATs with high delay spread channels. The pilot symbols are used for interference
estimation as well as to estimate the channel over every data symbol within a hop region using
time and frequency interpolation. Format 0 and Format 1 can support quasi-orthogonal RL and LS-
OFDMA transmissions (Section 1.3.12) with up to three and two overlapping users respectively.
The pilot pattern is indicated through the packet format signaling. The pilots and data symbols
within every block undergo the same transmit processing within each block.
- 35-
8 OFDM symbols 8 OFDM symbols
cluster
cluster
16 tones
16 tones
Format 0 18 pilots Format 1 24 pilots
SIMO + QORL (Q = 3) High delay spread
(SIMO + QORL (Q = 2)
Data symbol Pilot symbol
Figure 19: Reverse Link Pilot Formats
Pilot overhead for format 0 is 14.06 %, while that for format 1 is 18.75 %. Note that softer handoff
is supported with orthogonal pilot patterns (details TBD). The pilot patterns can be further
optimized based on evaluation results (TBD). The use of resource orthogonal pilots for QORL and
LS-OFDMA is under study (TBD). Subband hopping and diversity hopping can be multiplexed in a
frame. One more study item is to see if multiple interlaces can have common hopping for better
channel estimation (TBD).
1.2.1.2 Multiplexing Diversity and Subband Hopping
The RL allows quasi-static resource partitioning between diversity hopping and subband hopping
in units of subbands. The number of diversity hopping subbands is signaled via the overhead
channels. The logical subbands in diversity hopping mode are mapped to physical resources (tiles)
that are scattered across the entire bandwidth. The logical subbands in subband hopping mode are
mapped to physical subbands. The tiles within physical subbands that are punctured by diversity
hopping are replaced by available tiles outside the subbands. The number of base nodes per
subband is independent of resource partitioning, thus preserving the interference diversity order
irrespective of resource partitioning. The mapping of logical subbands in subband hopping mode to
physical subbands changes over time, thus ensuring that all physical subbands are available for
subband scheduling.
1.2.2 Reverse Link Channels
The RL channels of HKLLMNQRSUZ FDD are shown in Figure 20, and described in Table 6.
- 36-
PHY
R-CQICH R-REQCH R-ACKCH R-DCH R-CDCH
Channels
R-MCQICH R-BFCH R-PICH R-AuxPICH R-ACH
R-SFCH
Figure 20: RL Channels for HKLLMNQRSUZ FDD
R-PICH RL Pilot Channel
R-CQICH FL Channel Quality Indicator Channel
R-MCQICH FL MIMO Channel Quality Indicator Channel
R-AuxPICH RL Auxiliary Pilot Channel to support CDMA Traffic Channel etc.
R-BFCH Feedback Channel in Support of FL Precoding and SDMA
R-SFCH Feedback Channel in Support of FL Subband Scheduling
R-REQCH Request Channel, Used to Request RL resources
R-ACKCH RL Acknowledgement in Support of FL H-ARQ
R-ACH RL Access Channel
R-DCH RL OFDMA Traffic Channel
R-CDCH RL CDMA Traffic Channel
Table 6: Description of RL Channels for HKLLMNQRSUZ FDD
1.2.2.1 R-DCH
The traffic channel carrying data is called Reverse Data Channel (R-DCH). The R-DCH uses a rate
1/3 convolutional code for block lengths less than 128 and a rate 1/5 turbo code for block lengths
greater than 128. The code is punctured or repeated to achieve desired code rate. R-DCH uses
synchronous H-ARQ. The channel interleaver is based on a bit-reversal pattern providing almost-
regular puncture patterns and good interleaver distance properties at all code rates. The F-DCH
supports QPSK, 8PSK, 16QAM, and 64QAM - this allows a wide range of spectral efficiencies. To
avoid repetition of coded bits, the F-DCH uses modulation step-down at high spectral efficiencies,
and gains up to 1 dB for later transmissions. Table 7 lists packet formats for R-DCH.
- 37-
Packet Spectral Spectral Modulation order for each
efficiency on 2nd Max number of transmission
format efficiency on 1st
transmission transmissions
index transmission 1 2 3 4 5 6
0 0.25 6 2 2 2 2 2 2
1 0.50 6 2 2 2 2 2 2
2 1.0 6 2 2 2 2 2 2
3 1.5 6 3 2 2 2 2 2
4 2.0 6 3 3 2 2 2 2
5 2.67 6 4 4 3 3 3 3
6 4.0 6 4 4 3 3 3 3
7 Non-decodable 3.0 6 4 4 4 3 3 3
8 Non-decodable 4.0 6 4 4 4 4 4 3
9 4.0 6 6 6 4 4 4 4
10 5.0 6 6 6 4 4 4 4
11 6.0 3.0 6 6 6 4 4 4 4
12 Non-decodable 3.5 6 6 6 4 4 4 4
13 Non-decodable 4.0 6 6 6 6 4 4 4
14 Non-decodable 4.5 6 6 6 6 4 4 4
Table 7: Reverse Link Packet Formats
Packet formats 8 and 12-14 are non-decodable at the first transmission, but allow for high spectral
efficiency at target termination.
1.2.2.2 Reverse Link CDMA Traffic (R-CDCH)
Support for R-CDCH is optional for the AT. Each AT is assigned a CDMA control sub-segment and
may be assigned one or more CDMA traffic sub-segments that may overlap or be orthogonal to the
CDMA control sub-segment. CDMA access allows for a statistically multiplexed autonomous
transmission capability - useful for bursty low rate traffic as well as for delay-sensitive
applications. In addition, since there is no need for explicit scheduling grants, FL control signaling
overhead savings especially for cell-edge users can be easily realized.
CDMA capacity can be substantially enhanced with pilot and traffic interference cancellation.
CDMA also allows for fractional other sector interference cancellation. CDMA transmissions have
low PAPR, and hence are beneficial for power-limited users. If the network allows frame selection,
CDMA cell-edge users can further benefit from macro diversity gains.
- 38-
Fast performance based outer loop can be performed on the CDMA traffic segment at the AT. The
CDMA segment at each access point (AP) consists of multiple sub-segments, configured by the
network. If an AT supports CDMA traffic, the AN can assign any subset of these CDMA sub-
segments to the AT for data transmission. Full flexibility is available for assignment of the CDMA
sub-segment – it can be common across the AN and same for all ATs, or allow for partial overlap
across APs. It is also possible to have control sub-segments only (i.e., no traffic) for all ATs. Frames
transmitting R-CDCH also code multiplex Auxiliary pilot (R-AuxPICH) that is used as reference for
CDMA traffic demodulation. Hence R-AuxPICH occupies the same bandwidth as the data
transmission. The R-AuxPICH can also be used for control channel demodulation on frames where
data is transmitted. The control sub-segment hops over traffic sub-segments.
There are three packet formats for CDMA traffic that are optimized for VoIP traffic. Other types of
traffic may be transmitted on this segment subject to packet format limitation. The AN will indicate
which flows are allowed on the CDMA traffic segment only, OFDMA traffic segment only, or both. If
a flow is allowed on both the OFDMA and CDMA traffic segment, the AT will determine which
segment will be used by this flow.
Rate determination on the CDMA traffic channels is carried out using a Reverse Rate Indicator
(RRI) channel or by using blind detection. The need for RRI channel will be defined based on
complexity of blind rate determination through multiple decoding attempts at AP on one hand and
control overhead associated with RRI channel as well as RRI interference cancellation complexity
on the other hand. A supplementary F-ACKCH channel will be allocated within the FL control
channel segment to support ATs with enabled CDMA traffic.
1.2.2.3 Multiplexing CDMA and OFDMA Traffic
Error! Reference source not found.Error! Reference source not found.Figure 21 illustrates how
the CDMA and OFDMA traffic segments are multiplexed.
Figure 21: Multiplexing CDMA and OFDMA Traffic
Multiplexing of CDMA control and CDMA traffic can be flexibly configured. In the interlaces where
CDMA control is transmitted, CDMA traffic can be frequency multiplexed or code multiplexed.
Furthermore, CDMA traffic can be configured to use interlaces different than the one used by
CDMA control. This allows RoT operating points to be individually tuned for CDMA control and
CDMA data.
- 39-
1.2.2.4 Reverse Link Control Channels
The RL control channels include the acknowledgement channel (R-ACKCH), the pilot channel
(R-PICH) and supplemental broadband pilot channel, the channel quality indicator channel
(R-CQICH), the MIMO channel quality indicator channel (R-MCQICH), the auxiliary pilot channel
(R-AuxPICH), the request channel (R-REQCH), the access channel (R-ACH), the beamforming
feedback channel (R-BFCH), and the subband feedback channel (R-SFCH).
The control channels other than the acknowledgement channel are transmitted in a CDMA control
segment. The use of CDMA for the control segment provides statistical multiplexing benefits since
resources don‟t have to be reserved for all the channels. The R-CQICH also provides a broadband
reference for power control, and subband scheduling. Along with the R-PICH, R-CQICH provides a
broadband pilot that covers all the bandwidth over time. The presence of request and access
channels in the CDMA segment allows terminals to access the system and request resources with
low latency and minimal extra overhead. In a synchronous network, multiple sectors can have a
common CDMA control segment. This enables multiple sectors in the ATs active set to monitor its
control channel transmissions, enabling a fast and efficient handoff mechanism.
The CDMA control segment occupies a fraction of the RL bandwidth on one or more RL traffic
interlaces with the minimum allocation and granularity of one control subsegment. Each control
subsegment consists of 256 contiguous subcarriers for systems with 512 pt FFT and above, 128
contiguous subcarriers for systems with 256 pt FFT and 64 contiguous subcarriers for systems
with 128 pt FFT. This amounts to 1/12 bandwidth overhead for a 5 MHz system (see Figure 22).
The CDMA control segment occupies a multiple of control subsegments on any RL interlace that
carries a control segment and hops over the entire RL bandwidth. This allows for flexible load
control by changing persistence of different channels.
Figure 22: Reverse Link CDMA Control Segment
For each member of the active set, the AN assigns one primary control subsegment to the AT. The
AT should use this primary control subsegment to transmit any RL CDMA control channel, other
than the supplemental broadband pilot channel, to the corresponding member of the active set.
Preferably, the primary control subsegments assigned to a given AT overlap, so that AT does not
need to transmit multiple primary pilots on different control subsegments. To this end, after the
first subsegment is assigned at the time of access, the subsequent primary control subsegments
corresponding to the new members of the active set are assigned with AT assistance.
- 40-
Every control channel in the CDMA segment carries at most 10 information bits. Messages with a
smaller number of information bits will have the remaining bits set to „0‟ so that AN can make use
of this information to reduce receiver complexity and improve detection performance. Every 10-bit
message is mapped to a Walsh space of size 1024. Furthermore, various control channels are
scrambled with a pseudo-random scrambling which is defined by the channel type (e.g. R-CQICH),
target sector, and MACID, whenever applicable. The reporting rates and power offsets of each
control channel described are controlled on a per-AT basis using upper layer messages. The
number of control channels transmitted is adapted based on the rise over thermal and available.
Fast closed loop power control is used to set the transmit power levels on the RL control channels
that are transmitted periodically.
An AT can be served by different sectors on the FL and RL to achieve best cell site selection gains
in situations where the best serving sectors for FL and RL are different, usually referred to as
disjoint link scenario. In disjoint link scenarios, the AT transmits R-CQICH and R-MCQICH mainly
to the FL serving sector, and R-REQCH/R-AuxPICH to the RL serving sector. Also the R-ACKCH,
R-BFCH, and R-SFCH are transmitted only to the FL serving sector. In addition, as explained
later, R-PICH, R-MCQICH, and R-CQICH are transmitted to other members of the active set as
well, to enable RL channel quality measurement and power control of FL control channels.
In disjoint link scenarios, the fast power control is done by the RL serving sector. The transmission
power level of control transmissions to the FL serving sector is set based on a certain C/I target,
the PQI feedback, control rise over thermal (RoT) of the control segment, and traffic interference
over thermal (IoT) information from the FL serving sector.
1.2.2.4.1 Reverse Acknowledgement Channel (R-ACKCH)
The purpose of R-ACKCH is to acknowledge FL H-ARQ transmissions. For FL transmissions in
either SISO or single codeword MIMO mode, a single bit has to be transmitted per PHY frame in R-
ACKCH while multi-codeword MIMO transmissions will be acknowledged by one bit per layer.
An R-ACKCH ID is associated with every valid base node of the FL channel tree (see Figure 23).
This allows efficient support of H-ARQ with the maximum number of channel assignments (i.e.,
when all assignments have the minimum possible assignment size). A larger channel assignment
consists of a set of base nodes, each associated with an R-ACKCH ID. In this case, the R-ACKCH
ID associated with the base node with the lowest Channel ID is used. In MCW MIMO mode, the
acknowledgement message consists of multiple bits (one per layer). An MCW MIMO FL assignment
is restricted to include a number of base nodes at least as large as the number of MIMO layers. The
R-ACKCH IDs corresponding to different base nodes of such a MIMO assignment can be used to
acknowledge different MIMO layers. Some orthogonal dimensions are set aside for every R-ACKCH
ID as described below.
- 41-
Figure 23: R-ACKCH ID assignments
The R-ACKCH occupies Nt tiles that will subsequently be referred to as R-ACKCH tiles, over each
RL traffic interlace. Each tile is a collection of subtiles wherein each subtile spans 8 contiguous
subcarriers over 2 contiguous OFDM symbols. Each R-ACKCH tile occupies the lower 8 subcarriers
of some traffic tile. The number of subtiles per tile is four for 512pt FFT and above as shown in
Figure 24. The number of R-ACKCH tiles scales (with a granularity of one R-ACKCH tile) as
required by the number of traffic channels, with the minimum of four R-ACKCH tiles, if available,
in order to ensure channel and interference diversity of fourth order and a manageable link budget.
Each R-ACKCH subtile accommodates 8 R-ACKCH bits corresponding to 8 different traffic
channels. These bits are transmitted using on-off keying (OOK) and spread over the R-ACKCH
subtile with orthogonal spreading codes - a column of the DFT matrix of size 16. The remaining 8
codes of the DFT basis can be used by the AN for interference estimation. Each R-ACKCH ID is
transmitted over four R-ACKCH subtiles with every R-ACKCH subtile taken from a different R-
ACKCH tile, thereby ensuring 4-th order diversity. Furthermore, each R-ACKCH bit will be
multiplexed with different R-ACKCH bits on different R-ACKCH subtiles, in order to ensure some
diversity wrt. multiplexing interference that is due to some loss of orthogonality in time/frequency
selective channels. Finally, R-ACKCH tiles will be hopping randomly wrt. traffic tiles, to make sure
that R-ACKCH uniformly punctures different traffic channels. For systems with 128pt FFT and
256pt FFT, the nominal number of subtiles per tile is one and two respectively. Additional subtiles
will be added to some R-ACKCH tiles in a system with 128pt or 256pt FFT when the total number
of R-ACKCH tiles defined by the total available traffic tiles on a given interlace (which may be
reduced by the presence of RL CDMA control segment) does not allow for multiplexing all R-ACKCH
IDs. This situation may occur when SDMA is allowed on the FL. R-ACKCH structure for the cases
of 256pt and 128pt FFT are shown in Figure 25.
- 42-
8 OFDM symbols 8 OFDM symbols
8 codes for 8 R-ACKCHs
8 codes for interference
16 tones
16 tones
4- th order diversity
2
Subtile 4
Subtile 1
3
Subtile 2
4
Subtile 3
per R-ACKCH
Traffic
Traffic only &
R- ACKCH
Data symbol Pilot symbol
Figure 24: R-ACKCH structure: 512pt FFT and above
8 OFDM symbols 8 OFDM symbols
8 codes for 8 R-ACKCHs
8 codes for interference
-
16 tones
16 tones
4- th order diversity
Subtile 2
Subtile 3
Subtile 4
Subtile 1
per R-ACKCH
Traffic
Traffic only &
R- ACKCH
Data symbol Pilot symbol
Figure 25: R-ACKCH Structure: 256pt FFT and 128pt FFT
Every R-ACKCH ID occupies 8 complex dimensions while the total number of R-ACKCH IDs equals
to the number of (minimum) channels, with each channel occupies 16 subcarriers and 8 OFDM
symbols. Hence the equivalent overhead is 1/16. A CDMA ACK channel is under study (TBD).
1.2.2.4.2 Reverse Pilot Channel (R-PICH)
Each AT transmits a known pilot sequence on each control subsegment assigned to it. This pilot
sequence is scrambled with a scrambling sequence which is selected at the time of access, and is
made known to all new members of the active set. The scrambling sequence is a function of the PN
of the sector at which the AT accesses the AN, the MACID of the AT at that sector, and the access
time. This makes the scrambling sequence independent from AT‟s current serving sector, and
hence eliminates the need for backhaul or over the air communication with other members of the
active set to inform them of the new scrambling sequence after each handoff. It also ensures that
the uniqueness of the scrambling sequence is preserved as the AT moves around and hands off to
other sectors in the network.
- 43-
The RL serving sector of the AT uses the pilots on R-PICH to determine the RL channel quality from
the AT, and to determine the value of the power control command to that AT. Active set members
use this pilot to generate the pilot quality indicator (PQI) information for the AT, which is
transmitted to the AT on F-PQICH.
The AT uses the power control commands from its RL serving sector to adjust the transmit power
of the RL pilot channel. It also uses the PQI information for RL serving sector selection, and to set
the transmit powers on different RL control channels and RL traffic channel.
In addition to the primary pilot channel described above, the AT can transmit supplemental pilot
channels over a larger number of CDMA control subsegments assigned by its forward or RL serving
sectors. Each supplemental pilot channel is scrambled with the same scrambling sequence as the
primary R-PICH, and serves as a broadband reference to enable adaptive transmission such as RL
subband scheduling and FL beamforming.
1.2.2.4.3 Channel Quality Indicator Channel (R-CQICH)
The primary purpose of R-CQICH is to supply the AN with a FL channel quality measure that can
be used for scheduling transmissions on the F-DCH, and is frequently transmitted by every AT to
its FLserving sector, and cycles among the active set members. The R-CQICH includes a control
CQI report to support channel quality feedback for SISO transmission, and indicating the desired
FL serving sector for FL L1 handoff. The control CQI report can also be used by FL and RL serving
sectors and other members of the active set for power controlling the control information to the AT.
Examples of such control information are power control bits (F-PCCH), PQI information (F-PQICH),
and FL and RL assignment messages that are transmitted on F-SCCH.
The R-CQICH signals FL channel quality (4 bits) for control and SISO traffic across the band. It
also signals the desired FL serving sector indicator to request FL handoff. The R-CQICH is
scrambled with the scrambling sequence of its target sector.
1.2.2.4.4 MIMO Channel Quality Indicator Channel (R-MCQICH)
This channel carries the feedback for the MIMO channels on the FL. In SCW mode, the feedback
includes rank of the MIMO transmission suggested by the AT and CQI value which is
representative of the total spectral efficiency corresponding to this rank value. In MCW mode, the
feedback consists of CQI values per layer. Detailed structure of this feedback channel is under
study. (TBD)
1.2.2.4.5 Reverse Link Auxiliary Pilot Channel (R-AuxPICH)
The R-AuxPICH is used to provide demodulation reference for the CDMA traffic on the RL.
1.2.2.4.6 Request Channel (R-REQCH)
The primary purpose of R-REQCH is to request RL traffic resource allocation from the RL serving
sector. Any R-REQCH sent to a sector different from the RL serving sector is interpreted as RL
handoff request. The R-REQCH transmission includes indications of the QoS level associated with
the request, and the amount of resources requested, i.e. indicates the buffer level of the AT. The R-
- 44-
REQCH is scrambled with the scrambling sequence of its target sector. The exact details of the
message are under study (TBD).
1.2.2.4.7 Beamforming Feedback Channel (R-BFCH)
The R-BFCH provides feedback that allows for adaptive beamforming and spatial multiplexing
(SDMA) of multiple ATs on the FL. The detailed structure of this feedback channel is under study
(TBD).
1.2.2.4.8 Reverse Subband Feedback Channel (R-SFCH)
The R-SFCH provides feedback that allows for adaptive sub-band scheduling on the FL by
providing subband specific channel quality reports. The compact version of R-SFCH indicates the
preferred subband index. The full version of R-SFCH includes, in addition to the reported subband
index, the subband channel quality (CQI) value. The number of CQIs transmitted is adapted based
on the available TPR. Multiple CQIs can be transmitted in case TPR is available. (example: If RAB is
applied and the AT prefers to report them. The detailed structure of this feedback channel is under
study (TBD).
1.2.2.4.9 Reverse Access Channel (R-ACH)
The R-ACH is located in the CDMA control segment and carries the access preamble for initial
access and access based handoff. The R-ACH make use of preamble power ramping and achieves
90% access latency within 22ms and 99% access latency within 30ms.
The R-ACH is modulated as a 1024 Walsh sequence with the target sector scrambling. The entire
space of available probe sequences may be subdivided into a number of groups. The AT selects an
access probe sequence randomly from a group with the desired parameters such as buffer level,
measured FL strength etc. thereby communicating these parameters to the AN through the access
process. More details on the access procedure and logic can be found in Section 1.3.2. Unlike other
control channels within CDMA segment, the R-ACH has an extended guard band and guard time in
order to prevent intra-sector interference caused by a misalignment of the access probe with CDMA
segment boundaries resulting from the fact that AT in the access phase does not have accurate RL
timing information.
1.3 Operational Details of HKLLMNQRSUZ FDD
Some details of the control procedures are outlined in this section.
1.3.1 Acquisition
Each FL superframe consists of a superframe preamble followed by a sequence of FL frames (see
Figure 1 in Section 1.1.1). The superframe preamble consists of 8 OFDM symbols for FFT sizes of
512 and above. These symbols are indexed 0 through 7 as shown in Figure 26. The superframe
preamble consists of 16 OFDM symbols for the case of a 256 point FFT, and of 32 OFDM symbols
for the case of 128 point FFT.
For simplicity, the remaining description in Section 1.3.1 is restricted to FFT sizes of 512 and
above. For the other cases, each constituent channel of the superframe preamble is scaled
- 45-
(repeated in time) by the appropriate factor (factor of 2 in the case of a 256 point FFT and factor of
4 in the case of 128 point FFT) as compared to the case of a 512 point FFT. The preamble occupies
~5MHz (512 subcarriers less the guard tones) in any given superframe. In large deployments (>
5MHz), the preamble hops over the entire band, thus maintaining non-coherent processing
efficiency in multi-path channels by exploiting the gain due to random reuse.
F-PBCCH
Cyclic prefix
length,
Number of
guard sub-
carriers etc. F-SBCCH
Odd superframes:
- FL Hopping Structure TDM 1
- FL Pilot Structure
- FL Control Chanel Structure
+ TDM 2, 3
- Number of Transmit Antennas (F-OSICH)
- etc.
Even superframes:
- Quick pages.
OFDM Symbol Index (0-7)
Figure 26: Superframe Preamble Structure for FFT Sizes of 512 and Above
The last three OFDM symbols in the superframe preamble (the symbols indexed 5, 6 and 7) are
TDM pilots which are used for initial acquisition. These symbols will also be referred to TDM pilot
1, TDM pilot 2 and TDM pilot 3. The first of these forms the Acquisition Channel (F-ACQCH), while
the latter two are reused in order to transmit the Other Sector Interference Channel (F-OSICH).
TDM Pilot 1 is transmitted on OFDM symbol 5, while TDM Pilots 2 and 3 are transmitted on OFDM
symbols 6 and 7 respectively.
The structure of the last three OFDM symbols is as described in Sections 1.3.1.1 and 1.3.1.2.
The following items are under study: position and configurable length of F-PBCCH and F-SBCCH in
the preamble, as well as the preamble structure (frequency domain mapping of sequences to
antennas, number of preamble symbols). The details are TBD.
1.3.1.1 TDM Pilot 1
TDM Pilot 1 consists of one OFDM symbol, and is fixed across the deployment. The construction of
TDM Pilot 1 depends on whether the bandwidth is less than 5 MHz or greater than 5 MHz.
The TDM1 OFDM symbol is constructed using a Generalized Chirp Like (GCL) sequences, see
Section 4.3 for more details. This is followed by an IFFT and the addition of cyclic prefix and
windowing intervals, like in all other OFDM symbols. The GCL sequence used for this purpose
does not depend on the sector, therefore it is not possible to identify sectors using TDM Pilot 1.
In a bandwidth of more than 5 MHz, the bandwidth is split into segments which have a bandwidth
of 5 MHz each. A sector transmits a TDM Pilot 1 only on one segment in any given superframe.
There is one option for FFT sizes of 128, 256, and 512, two options for FFT sizes of 1024, and four
options for FFT size of 2048. However, the segment on which the sector transmits the pilot varies
- 46-
from superframe to superframe. The segment is chosen differently in the two different
synchronization modes, as described in section 1.3.1.3. The time domain sequence used for
modulation is again independent of the sector and depends only on the total bandwidth and the
index of the segment. The design for bandwidths less than 5MHz is TBD.
However, the GCL sequence conveys the following information: System FFT size (128, 256, 512,
1024, or 2048), index of a 5MHz carrier on which this TDM1 is populated for bandwidths exceeding
5MHz, one bit indicating whether frequency reuse on PBCCH and SBCCH is used, and more
information may be incorporated to assist in system determination (TBD). A total of TBD sequences
are needed.
TDM Pilot 1 can be used to get an estimate of the superframe timing. Once a timing is identified,
the AT can proceed to sector identification using TDM Pilots 2 and 3. TDM Pilot 1 may also be
used for frequency correction.
1.3.1.2 TDM Pilots 2 and 3
TDM Pilots 2 and 3 are used for sector identification after one or more strong paths have been
acquired using TDM Pilot 1. TDM Pilots 2 and 3 are constructed using a time-domain sequence
that is one of a set of 1024 Walsh sequences scrambled by a PN sequence. The first 512 elements
of the Walsh sequence are used to construct TDM Pilot 2 and the latter 512 to construct TDM Pilot
3. The Walsh sequence depends only on the sector‟s PilotPN and does not change from superframe
to superframe. Out of the 1024 sequences, 512 sequences are reserved for Semi Synchronous
mode and the other 512 are reserved for Asynchronous mode. The PN sequence depends only on
the system bandwidth in the case of a deployment with bandwidth ≤ 5 MHz and on the system
bandwidth and segment index in case of a deployment with bandwidth > 5 MHz. The segment on
which TDM Pilots 2 and 3 are transmitted is identical to the segment on which TDM Pilot 1 is
transmitted and is determined as described in section 1.3.1.3.
Use of GCL sequences for TDM2 and TDM3 is being investigated (TBD).
1.3.1.3 Synchronization Modes
The system supports two modes, namely Semi-synchronous and Asynchronous, and the TDM
pilots are generated differently in the two cases.
1.3.1.3.1 Semi Synchronous Mode
For some applications, such as ranging, it is necessary to detect extremely weak sectors. For this
reason it can be important to allow correlations over more than one superframe preamble. In order
to get processing gains over more than one superframe preamble, it is necessary that the
interfering signal (i.e., the TDM pilot of the neighboring sector) changes from one superframe
preamble to the next. To enable this, an auxiliary quantity that changes from superframe to
superframe called the PilotPhase is defined for each sector. The PilotPhase is a 9 bit quantity
defined as PilotPhase = (PilotPN + SuperframeIndex) mod 512, where SuperframeIndex is a counter
that is incremented from one superframe to the next, and is defined globally across all sectors. The
PilotPhase is used to generate the scrambling sequences for the two TDM pilots. Furthermore, the
segment on which the TDM Pilots are transmitted in a deployment with BW > 5 MHz is chosen
pseudo-randomly as a function of PilotPhase.
- 47-
This pilot structure requires some level of synchronization between two sectors, as different sectors
use offsets of the same sequence. To be more precise, if all possible values of PilotPN are possible,
then this structure requires that any two sectors be synchronized to within half a superframe – but
not at the symbol or chip level. Otherwise it is conceivable that two sectors with different PilotPNs
will transmit the same acquisition pilots (same PN sequences) at the same time. This mode can be
used to improve performance (reduce acquisition time, fast sector switching, interference
estimation etc).
1.3.1.3.2 Asynchronous Mode
In some situations it is not possible to accurately synchronize two sectors. In order to support
these scenarios, there is a mode in the system that has no synchronization requirements. In this
mode, the TDM pilots are scrambled directly using the PilotPN instead of the auxiliary quantity
PilotPhase. For a system with bandwidth ≤ 5 MHz, the TDM pilots are the same from superframe to
superframe. For a system with bandwidth > 5 MHz, the TDM Pilots cycle deterministically through
the set of available segments.
1.3.1.4 Acquisition Procedure
The first five OFDM symbols in the superframe preamble are used to carry the two Primary
Broadcast Channels, namely F-PBCCH and F-SBCCH. These channels carry configuration
information that the AT needs to have before it can demodulate the PHY Frames. In addition, the
F-SBCCH channel also carries paging information.
An F-PBCCH packet is encoded over 16 superframes, and occupies ¼ of an OFDM symbol in each
superframe preamble - an extremely small overhead. F-PBCCH carries deployment-wide static
parameters like cyclic prefix duration, number of guard subcarriers, in addition to the superframe
index, and is required only at initial wake-up.
An F-SBCCH packet is encoded over a single superframe and occupies 4 ¾ OFDM symbols in each
superframe preamble. The bandwidth overhead of this channel is approximately 2%. F-SBCCH
carries sufficient information to enable the AT to demodulate FL data from the PHY Frames like
information on FL resource channels and multiplexing modes, pilot structure, control channel
structure, transmit antennae, etc. This information is transmitted every alternate superframe; the
other superframes are used to carry pages The remaining overhead information is broadcast using
a regular data channel in predefined superframes – this carries information on RL hopping
patterns, channel mapping, transmit powers, power control parameters, access parameters, etc.
These channels enable a flexible physical layer, allowing a flexible configuration of cyclic prefix,
number of antennae, pilot structure, etc. They can also support FL and RL control channels with
flexible overheads, which can be matched to the current user loads. Also, features like sub-band
scheduling, FFR etc. can be enabled or disabled.
On initial wake-up, the AT first detects a sector and achieves time and frequency synchronization
using the TDM pilots. In SemiSynchronous mode, the AT knows the value of the PilotPhase
variable at the end of this stage, while in Asynchronous mode, the AT knows the value of the
PilotPN variable at the end of this stage. The AT then goes on to demodulate the F-PBCCH and F-
SBCCH channels. The F-PBCCH channel carries the lower 9 bits of the SuperframeIndex, which
- 48-
enables the AT to find the value of PilotPN in the SemiSynchronous case (PilotPN = PilotPhase –
SuperframeIndex mod 512). Therefore, in both SemiSynchronous as well as Asynchronous modes,
at the end of this stage, the AT knows the PilotPN and SuperframeIndex variables, which are
together used to seed various random number generators (for hopping, scrambling etc) used in
generating the FL waveform.
The Other Sector Interference Channel (F-OSICH) carries a three state quantity that is modulated
as a phase on TDM pilots 2 and 3. Since the TDM pilot waveform is known once acquisition is
completed, the superposition causes no degradation to the performance of the OSICH. The function
of this channel will be described in detail in Section 1.3.7.4. One of the characteristics is that it is
used by ATs in the neighboring sector, i.e., it should be decodable at extremely low SNRs. This is
accomplished by providing an extremely large spreading gain for this channel, i.e., an entire OFDM
symbols is used to transmit less than two bits of information.
Frequency reuse is enabled on F-PBCCH and F-SBCCH as a deployment-wide feature for sectors in
semi-synchronous acquisition mode – this feature is optional at the AN and mandatory at the AT.
Frequency reuse is implemented by defining orthogonal hopping sequences associated with sector
IDs, hence allowing for frequency planning. Frequency reuse on F-PBCCH, F-SBCCH and SFN
paging are not supported simultaneously. Frequency reuse is not used in the asynchronous
acquisition mode of operation. Note that the presence of frequency reuse is used as a part of
system determination - the number of possible reuse schemes needs to be limited. The current
working assumption is a reuse of 1/7 – the details are TBD.
1.3.2 Paging
Paging is the process by which the AN initiates a connection with an AT that is in idle state, such
that the AT wakes to listen to the FL traffic only at certain negotiated time intervals. Page is
delivered by using QuickPages on the Control Channel and by using Page message on the Forward
Traffic Channel. To enhance the performance of paging, some features (e.g. fast repage) will be
considered in Working Group 2.
1.3.2.1 SFN Operation
The quick paging channel is transmitted every even superframe on F-SBCCH. The quick page may
be transmitted in SFN by sectors belonging to the same quick paging group. It is accompanied by
SFN transmission of broadband pilot to provide better F-QPCH performance at the cell edges by
helping in interference reduction in interference limited scenarios. The scheme gets full diversity
advantage in both slow and fast fading.
SFN quick paging is optional at the AN and mandatory at the AT. Regular pages are carried on the
F-DCH, and can be sent in an SFN mode. The AN can reserve resources in the adjacent sectors
and/or serve pages through softer handoff groups.
1.3.3 Access Channel Procedures
The access channel (R-ACH) is used by the AT to gain initial access to the network. The access
succeeds when the AT receives an acknowledgement sent by the AN upon successful detection of
an access probe. The access probes are multiplexed with other RL control channels in a CDMA
control channel segment.
- 49-
On initial wake-up, the AT first detects a sector and achieves time and frequency synchronization,
finds the PilotPN and SuperFrameIndex variables, and the complete configuration of the FL and
RL. The AT accesses the system by sending successive access probes on the CDMA control
segment. The access succeeds when the AT receives an AccessGrant message sent by the AN over
the F-SCCH upon successful detection of an access probe. The AccessGrant message assigns to the
AT a MACID and initial RL resources, and a timing adjustment command so that the AT can
properly orthogonalize its transmission to all the other ATs.
1.3.3.1 Access Probe Structure
The Access Channel MAC Protocol transmits access probes by instructing the PL to transmit a
probe including the power level, AccessSequenceID, PilotPN of the sector to which the access probe
is to be transmitted among other parameters. An access probe is transmitted over the CDMA
control segment and it is modulated as a 1024 Walsh sequence (given by the AccessSequenceID)
scrambled with a PN sequence generated using the PilotPhase of the FL Serving Sector and the RL
PHYFrame as seeds.
Some information is embedded in each access probe through the selection of the Walsh sequence.
The space of the Walsh sequences is partitioned into a number of sets, each set corresponding to a
possible value of the variable to be transmitted. For each access probe, the Walsh sequence is
chosen randomly out of the sequences of the appropriate set. This method is used to send
information about the received pilot level and to request different amount of resources for
subsequent transmissions.
1.3.3.2 Access Probes Transmission Procedure
Figure 27 illustrates the grouping of access probes in access probe sequences. In the figure, Ns
probe sequences are shown, where each probe sequence has Np probes. The number of slots left
empty between successive probes from the same access sequence is given by AccessCycleDuration.
The term “slot” denotes a RL PHYFrame which contains the CDMA segment.
The probes of each sequence are transmitted at increased power. The ramping of power of
successive access probes in one sequence is designed such that most of the time an AccessGrant is
received before the complete transmission of the first access sequence (maximum Np probes). If
after the transmission of one access sequence no AccessGrant is received, the AT starts the
transmission of a new access sequence after waiting a random number of slots.
The number of slots before the retransmission of a new access sequence is determined in the same
way as the number of slots before the transmission of the first access probe - this random number
is referred to as the persistence interval which can be determined depending on cases.
The transmission of access probes stops (without completion of the current access sequence) if any
of the following conditions are met:
The AT receives an AccessGrant, or
Transmission is aborted because the MAC protocol at AT received a Deactivate command.
- 50-
AccessCycleDuration AccessCycleDuration AccessCycleDuration
p-persistence
p-persistence
p-persistence
...
probe 1 2 3 Np 1 2 3 Np 1 2 3 Np
Time
probe sequence 1 2 ... Ns
Figure 27: Access Probe Sequences. Ns Sequences with Np Probes per Sequence
For each access probe, the MAC protocol chooses a different AccessSequenceID randomly as
described before. This reduces the probability of successive collisions between the access probes of
two different ATs.
1.3.4 Timeline
Both FL and RL data transmissions support H-ARQ. To provide H-ARQ related processing time at
the AN and AT, an eight interlace structure is used for both the FL and RL. Other interlace
structures are also provided as possible alternatives.
1.3.4.1 Forward Link Timeline
The eight interlace structure is accepted as the baseline, as it provides a relaxed processing
timeline at the AT. Timing of transmissions associated with one of the eight interlaces is shown for
FL in Figure 28. Other interlaces are shifted by the same number of frames. This interlace
structure ignores the presence of the superframe preamble. Acknowledgment is sent at a delay of
five frames and retransmission occurs after eight frames. H-ARQ retransmission latency is
approximately 7.3 ms.
0 1 2 3 4 5 6 7 8
Assignment Data (Re-)
AN Data Transmission Transmission
ACK
Transmission
AT
PHY Frame (TTI) Demod + Decode Time ACK Decode + Scheduling + Encoding Time
8 OFDM Symbols 32 OFDM Symbols 16 OFDM Symbols
Retransmission Interval
64 OFDM Symbols
(Approximately 7.3ms)
Figure 28: Forward Link H-ARQ Timeline
The Extended Frame timeline shown in Figure 29 allows for packet encoding over three frames.
Note that Extended Frame timeline has the same retransmission latency as the regular timeline,
- 51-
namely eight frames. Hence, Extended Frame interlaces can coexist with the regular (single frame)
interlaces. Examples of such a coexistence are a single Extended Frame interlace followed by five
regular interlaces and two Extended Frame interlaces followed by two single frame interlaces.
Extended Frame interlaces can be useful when small assignments and low spectral efficiencies
result in small packets and therefore relatively high MAC header overhead and small coding gain.
Since low spectral efficiencies can be avoided on FLby assigning a higher share of the transmit
power to users with weak channel conditions, the need for Extended Frame transmission needs to
be further evaluated.
0 1 2 3 4 5 6 7 8
Assignment
Data Transmission Data Transmission Data Transmission Data (Re-)
AN Transmission
Extended Frame
ACK
Transmission
AT
PHY Frame (TTI) Demod + Decode Time ACK Decode + Scheduling + Encoding Time
24 OFDM Symbols 16 OFDM Symbols 16 OFDM Symbols
Retransmission Interval
64 OFDM Symbols
(Approximately 7.3ms)
Figure 29: Forward Link H-ARQ Timeline for Extended Frame Transmission
The following timelines are also under study: The six interlaces is shown for FL in Figure 30 - the
other interlaces are shifted by the same number of frames. This interlace structure ignores the
presence of the superframe preamble. Acknowledgment is sent at a delay of three frames and
retransmission occurs after six frames. Assuming a processing delay of two frames, the H-ARQ
retransmission latency is approximately 5.5 ms.
Figure 30: Alternative Forward Link H-ARQ Timeline: Six Interlaces
- 52-
A five interlace structure as described in Figure 31 is also under study.
Figure 31: Alternate Forward Link H-ARQ Timeline: Five Interlaces
As explained above, in addition to the one-slot frame, a new frame structure will also be added. The
same signaling mechanism that is used to configure the one-slot frames will be extended for the
new frame structure – the exact details are TBD. For example, in a new structure, a single
transmission could consist of two adjacent PHY frames. The retransmission interval would be some
multiple of these two PHY Frames, so as to allow sufficient processing time.
1.3.4.2 Reverse Link Timeline
As a baseline, an eight interlace structure is used to provide a relaxed processing timeline at the AT
and AN. Timing of transmissions associated with one of the eight interlaces is shown for RL in
Figure 32. Other interlaces are shifted by the same number of frames. This interlace structure
ignores the presence of the superframe preamble. Acknowledgment is sent at a delay of five frames
and retransmission occurs after eight frames. H-ARQ retransmission latency is close to 7.3 ms.
0 1 2 3 4 5 6 7 8 9 10 11
Assignment
Assignment
AN ACK
Transmission
Data Data (Re-)
Transmissio Transmission
AT n
PHY Frame (TTI) Assignment Decode + Encoding Time Demod + Decode + Scheduling + SSCH Encoding Time Assignment / ACK Decode + Encoding Time
8 OFDM Symbols 16 OFDM Symbols 32 OFDM Symbols 16 OFDM Symbols
Retransmission Interval
64 OFDM Symbols
(Approximately 7.3ms)
Figure 32: Reverse Link H-ARQ Timeline
The Extended Frame timeline shown in Figure 33 allows for packet encoding over three frames.
Note that Extended Frame timeline has the same retransmission latency as the regular timeline,
namely eight frames. Hence, Extended Frame interlaces can coexist with the regular (single frame)
interlaces. Examples of such a coexistence are a single Extended Frame interlace followed by five
- 53-
regular interlaces and two Extended Frame interlaces followed by two single frame interlaces.
Extended Frame interlaces can be useful when small assignments and low spectral efficiencies
result in small packets and therefore relatively high MAC header overhead and small coding gain.
Note that Extended Frame transmission is instrumental in handling power limited users with high
QoS traffic and therefore is mandatory to ensure acceptable performance for e.g. VoIP user located
at the edge of a relatively large cell.
0 1 2 3 4 5 6 7 8 9 10 11
Assignment
Assignment
AN ACK
Transmission
Extended Frame
Data Data Data Data (Re-)
Transmission Transmission Transmission Transmission
AT
PHY Frame (TTI) Assignment Decode + Encoding Time Assignment / ACK Decode + Encoding Time
8 OFDM Symbols 16 OFDM Symbols 16 OFDM Symbols
Demod + Decode + Scheduling + SSCH Encoding Time
32 OFDM Symbols
Retransmission Interval
64 OFDM Symbols
(Approximately 7.3ms)
Figure 33: Reverse Link H-ARQ Timeline for Extended Frame Transmission
As an alternative to provide a faster turn-around time, a six interlace structure is being considered
as shown in Figure 34 for the RL – one interlace is shown, and the other interlaces are shifted by
the same number of PHY frames. This interlace structure ignores the elongation of the first frame.
The data transmission occurs two frames after the assignment message is sent, and is
acknowledged four frames after the transmission. H-ARQ retransmissions occur six frames later,
providing a retransmission latency of 5.5 ms with 0.9 ms (1 frame) of processing time at the AT,
and 2.7 ms (3 frames) of processing time at the AN.
0 1 2 3 4 5 6 7 8 9
Assignment,
Assignment
AN Transmission
ACK
Transmission
Data
Data
AT Transmission
(Re)Transmissi
on
Frame Duration (TTI) Assignment Decode Demod + Decode + Scheduling Assignment / ACK Decode
8 OFDM Symbols + Data Encode Time + SSCH Encoding + Data Encode Time
16 OFDM Symbols 16 OFDM Symbols 16 OFDM Symbols
Retransmission Interval
48 OFDM Symbols
(Approximately 5.5ms)
Figure 34: Alternative Reverse Link H-ARQ Timeline: Six Interlaces
The corresponding five interlace structure for the RL is shown in Figure 35.
- 54-
Figure 35: Alternative Reverse Link H-ARQ Timeline: Five Interlaces
Other structures are under study as noted for the FL. The trade-off is between a quick turn around
time vs. the processing power required at the AT and AN to be able to meet the timing deadline for
high data rates.
HKLLMNQRSUZ also allows for a duplexer free FDD mode supports operation in FDD spectrum
with reduced spacing between the FL and RL. The timeline for this mode is described next.
1.3.4.3 Half Duplex Mode
The half duplex mode makes use of fragmented spectrum, where the two transmit bands are close
enough that the duplexer for FDD operation may become prohibitively expensive. The ATs using
this mode could be low-cost multi-band terminals that transmit and receive in different spectrum
(FDD operation). Each AT only transmits or receives in any given slot, i.e. will have a duty cycle for
both the receive and transmit chains.
The timeline consists of two “half-duplex interlaces”, with each half-duplex interlace consisting of
alternating FL and RL PHY Frames. An FL PHY Frame in one half-duplex interlace corresponds to
an RL PHY Frame in the other half-duplex interlace. Both the half-duplex interlaces share a
superframe preamble. The RL CDMA control segment is replicated on both half-duplex interlaces.
The half-duplex ATs can use the control segment on only one interlace, while a full-duplex terminal
can use the control segment on either interlace.
The relationship between the assignments, data transmission, acknowledgement and
retransmission are as shown in Figure 36 and Figure 37 for the FL and RL respectively.
0 0’ 1 1’ 2 2’ 3 3’ 4 4’
Assignment Assignment Data (Re-) Data (Re-)
AN Data Transmission Data Transmission Transmission Transmission
ACK ACK
Transmission Transmission
AT
PHY Frame (TTI) Demod + Decode Time ACK Decode + Scheduling + Encoding Time
8 OFDM Symbols 32 OFDM Symbols 16 OFDM Symbols
Retransmission Interval
64 OFDM Symbols
(Approximately 7.3ms)
Figure 36: Forward Link H-ARQ Timeline for Half Duplex Mode
- 55-
Assignment Assignment
Assignment Assignment
AN ACK ACK
Transmission Transmission
0 0’ 1 1’ 2 2’ 3 3’ 4 4’
Data Data Data (Re-) Data (Re-)
Transmission Transmission Transmission Transmission
AT
PHY Frame (TTI) Assignment Decode + Encoding Time Demod + Decode + Scheduling + SSCH Encoding Time Assignment / ACK Decode + Encoding Time
8 OFDM Symbols 16 OFDM Symbols 32 OFDM Symbols 16 OFDM Symbols
Retransmission Interval
64 OFDM Symbols
(Approximately 7.3ms)
Figure 37: Reverse Link H-ARQ Timeline for Half Duplex Mode
The above structure is a eight-interlace structure similar to the eight-interlace FDD timeline, with
each half-duplex interlace consisting of four interlaces.
The AT randomly picks one of the half-duplex interlaces for access. Subsequently, the choice of the
half-duplex interlace is defined by the assigned MACID. Odd MACIDs correspond to one half-
duplex interlace, while even MACIDs correspond to the other half-duplex interlace. PC bits for a
given MACID are sent only on the corresponding half-duplex interlace. The AN can assign
resources to a half-duplex terminal on the half-duplex interlace corresponding to its MACID. The
AN can assign resources to a full-duplex terminal on any PHY Frame.
Note that these timelines applies for OFDM data transmissions only, the retransmission timeline
for CDMA systems are still TBD.
1.3.5 Resource Management
This section discusses how resources are allocated to the FL and RL data channels. Resource
allocation including rate determination is centralized at the AP for both FL and RLs. For the FL,
resource allocation is based on the FL channel quality reports from AT. For the RL, resource
allocation is based on measurements of the RL channel quality as well as RL feedback from the AT
including resource requests. The AN assigns FL and RL resources to the AT via the Shared Control
Channel (F-SCCH). CDMA traffic resource allocation is AT-centric (i.e. autonomous) once the
CDMA traffic zone has been allocated by the AP.
1.3.6 Scheduling
Scheduling refers to the allocation of subcarriers and spectral efficiency 4 to ATs over time and is
centralized in the AN for both FL and RL allocations to ensure orthogonal allocations of resources
to different ATs in the system. The scheduler aims to maximize system capacity while managing
QoS requirements such as latency and throughput requirements of ATs. Additionally, the
scheduler manages fairness across ATs that can have widely disparate link qualities and thus can
support different instantaneous spectral efficiencies.
4 Spectral efficiency is defined by coding and modulation packet formats (cf. Section 4.2).
- 56-
Although the scheduler design is implementation dependent, the design ensures that the scheduler
has information required to utilize features such as sub-band scheduling, fractional frequency
reuse, precoding, and SDMA to achieve the above goals. The CDMA traffic zone transmissions allow
for bursty low bit rate traffic and cell-edge users to communicate without explicitly being scheduled
by the AN. The CDMA resources are allocated in a semi-static fashion by the AN. Fast PC enables
precise QoS control and slower rate/fairness control can be enabled with a slow RAB bit.
1.3.7 Assignment Management
The HKLLMNQRSUZ FDD system has been designed to efficiently support a range of applications,
with a corresponding range of required throughput, latency, and application packet sizes. For
example, the system has been designed to support very high best-effort data capacity as well as
supporting many users of voice-like applications. In addition the system has been designed to
efficiently support a combination of disparate application types simultaneously over the air
interface.
Resources in the system are assigned in units of logical subcarrier, where a logical subcarrier is a
static resource maps to a physical subcarrier. Frequency hopping is implemented by having time
varying mappings from logical subcarriers to physical subcarriers. To reduce assignment signaling
overhead, the system uses “synchronous H-ARQ” and provides support for persistent assignments.
With synchronous H-ARQ, the resources for successive retransmissions are not independently
scheduled, but rather are retained for all retransmissions associated with a packet. Thus, the
assignment of a set of logical subcarriers applies to an “interlace”. Assignments on different
interlaces are independent, and an AT may be given resources on multiple interlaces. To provide
flexibility in resource multiplexing and improve resource utilization, resource adaptive synchronous
HARQ (RAS-HARQ) is used where the resource assigned to a data packet can be changed during
retransmission. Details are given in 1.3.7.2.
Assignments can be persistent or non-persistent. When an assignment is non- persistent, the
assignment expires on successful packet decode, or when the packet fails to decode after the
maximum number of H-ARQ retransmissions allowed for the packet. When assignments are
persistent, the assignment persists as long as the assigned resource is in use or alternatively when
the assigned resource is not de-allocated by the AN using an explicit message. For the former, an
assignment is in use as long as either a packet or an erasure sequence is transmitted using the
assignment. The erasure sequence is simply a one-bit “keep alive” indication used to inform the
receiver that the assignment should be retained even though a data packet might not be available
for transmission using the assignment. If neither a packet nor an erasure sequence is transmitted
using the assignment, the assignment expires and the resources are free for subsequent allocation.
The persistent assignment also expires as a result of packet failure(s). For the latter, the
assignment persists as long as it is not de-allocated. The persistent resource can be temporarily re-
assigned in a non-persistent fashion to other users. For both cases, it is possible for the AN to send
an explicit message that expires an assignment, i.e., persistent assignments persist until
supplemented, decremented or de-assigned. Persistent assignments are useful to reduce
assignment overhead required when it is beneficial to schedule multiple users simultaneously, and
to eliminate request latency for RL transmissions.
- 57-
1.3.7.1 Forward Link Channelization
Two types of assignments, i.e. BRCH and DRCH are used for the FL, and may both co-exist in a FL
frame. Details are described in 1.1.3.
1.3.7.1.1 Resource Allocation
To reduce the overhead required to specify sets of logical subcarriers in the system, a finite space
of channel IDs are defined that map to specific sets of logical subcarriers and that are used to
communicate assignments to ATs. Because assignments can be persistent, and to combat
fragmentation of resources in the system due to this finite mapping of channel IDs, the system
supports supplemental assignments that either increment or decrement the existing persistent
assignment. Such supplemental assignments are sent to augment an ATs allocation between
packet transmissions.
The mapping between channel IDs and logical subcarriers is defined using a channel tree, such as
the one illustrated in Figure 38. Each node on the tree is given a unique channel ID. Further, each
base node is mapped to a set of logical subcarriers. A channel ID completes the mapping to the set
of logical subcarriers.
Figure 38: Example Channel Tree
FL and RL resource assignments are communicated to the ATs via link assignment messages
transmitted over the F-SCCH. For details on the FL control structure see Section 1.1.5.3.
1.3.7.2 Resource Adaptive H-ARQ (RAS H-ARQ)
The H-ARQ transmissions follow the timeline outlined in Section 1.3.4, i.e. synchronous H-ARQ is
performed. If necessary, the resource assigned to a data packet can be changed starting at the next
H-ARQ transmission for retransmission using an assignment message. For non-persistent
assignment, assignment message is sent for the 1st sub-packet transmission. Assignment message
can be sent for subsequent sub-packets to change the resource assignment
For persistent assignment, assignment message is sent for the initial persistent assignment.
Assignment message can be sent for retransmission sub-packets to change the resource
assignment for retransmission sub-packets. Resources persistently assigned to an AT can be
- 58-
temporarily assigned to a different AT using non-persistent assignment message without de-
assignment of the persistent assignment, e.g., in case of H-ARQ early termination (details TBD).
Assignment messages can also be sent to override the current persistent assignment.
1.3.7.3 Dynamic Resource Sharing between Forward Link Control and Traffic
As explained in section 1.1.5.3, FL control segment is present in every FL frame and is allocated in
units of base nodes with the minimum of three nodes and increments of one node resulting in a
minimum allocation of 10% and granularity of 3.3%. Note that FL control segment includes F-
SCCH which carries various types of resource control messages as described in section 1.1.5.3.1. A
quasi-static allocation of FL control segment is dimensioned for the maximum anticipated F-SCCH
load and is signaled through FL overhead channel (F-SBCCH). In the presence of control load
variations, a variable fraction of the control resources can remain unused in each frame. The
dynamic sharing approach described in this section allows for a reuse of wasted control resources
by the data channel (F-DCH) on a frame by frame basis.
First, channel nodes allocated to FL control segment are addressable through the FL traffic
resource channel tree and can therefore be assigned to an AT in addition to regular F-DCH nodes.
Unlike regular F-DCH assignments, the use of FL control nodes to carry traffic symbols is
conditioned on the presence of one or more unused control resources.. Whenever one or more
control nodes are assigned to an AT, all modulation symbols associated with these nodes that do
not carry any F-SCCH messages in a given frame will contain data modulation symbols for that AT
in the said frame. Portions of the FL control nodes that are thus used to carry data modulation
symbols are referred to as supplementary F-DCH resources.
In order to support this feature, the status of F-SCCH utilization needs to be conveyed to all ATs
that are assigned control nodes, on a frame by frame basis. To this end, AP can transmit a Packet
Data Control Assignment Message (PDCAM) that carries F-SCCH resource utilization bitmask in
the FL control segment of every frame. The size of F-SCCH utilization bitmask is variable (with a
minimum length of 3), and is determined by the maximum number of potentially unused F-SCCH
messages according to the quasi-static allocation of FL control segment while every bit of this mask
indicates utilization status of the associated F-SCCH message. Thus, an AT that is assigned FL
control nodes and receives PDCAM will be able to determine which modulation symbols associated
to FL control nodes within its assignment are not used by F-SCCH and therefore carry data
symbols. The detailed format of the utilization bitmask and channelization of PDCAM is under
study. A candidate solution is to define PDCAM as an additional resource management message
type, along with the existing message types in F-SCCH as described in section 1.1.5.3.1.
Capacity gains due to dynamic resource sharing are highly dependent on traffic scenario, and in
particular the number of FL control nodes and the inactivity factor of each F-SCCH resource. Given
an additional overhead associated with the transmission of PDCAM and the traffic-dependent
nature of the gains due to dynamic sharing, the support of this feature is optional at the AN.
Likewise the support of this feature is optional for AT. An AT that does not support the dynamic
sharing feature shall not be assigned FL control nodes by the AN.
Note that power allocation for F-SCCH utilization bitmask transmission in a given frame should
target all ATs that are assigned FL control nodes and are sent some data modulation symbols on
these resources in the said frame. In order to minimize the cost of bitmask transmission, AP
- 59-
should strive to assign these nodes to ATs with a relatively high geometry. However the dynamics
of resource utilization in the presence of mixed traffic and ATs in different channel conditions will
make transmission of F-SCCH utilization bitmask to the weakest AT with control nodes within its
assignment too expensive in some scenarios.
An AT that is assigned some FL control nodes and does not detect F-SCCH utilization bitmask in a
given frame should be able to successfully demodulate the portion of data symbols sent on the
regular F-DCH resources. To this end, data modulation should be carried out in such a way that
the set of modulation symbols sent over the regular portion of AT‟s assignment should not depend
on the availability of FL control resources. A modulation strategy that achieves this goal is
illustrated in Figure 39. First, a circular buffer is filled out clockwise with the entire sequence of
coded bits coming out of the channel interleaver. Modulation of the regular F-DCH resources
across H-ARQ transmissions will be accomplished by reading out portions of coded bits from the
circular buffer clockwise and mapping these bits to the regular F-DCH resources according to the
packet and pilot format of the AT‟s F-DCH assignment. Note that this rule is consistent with the
data modulation principle when dynamic sharing is not enabled. On the other hand, modulation of
the supplementary F-DCH resources (i.e., portions of control nodes currently not used by F-SCCH)
will be accomplished by reading out portions of coded bits from the circular buffer
counterclockwise and mapping these bits to the modulation symbols available within FL control
nodes assigned to the AT. Modulation format to be used on the supplementary F-DCH resources is
under study. Not only does the described modulation strategy allow for unambiguous
demodulation of the regular F-DCH resources whenever supplementary F-DCH resources are also
present within the assignment, it also guarantees the use of all coded bits produced by the encoder
before any of coded bits is repeated.
- 60-
Counter clock-wise Clock-wise starting
starting point of the first point of sub-packet 0
set of “extra” encoder
bits due to inactive F- and 1st bit of encoder
SCCH resources output
Counter clock-wise
starting point of the next
set of “extra” encoder bits
due to inactive F-SCCH
resources
Last bit of
encoder
output
Starting point of
sub-packet 1
Figure 39: Data Modulation in the Presence of Dynamic Sharing
1.3.7.4 Study Items for Assignment
The following items are TBD. Shared persistent assignment of multiple ATs to receive the same FL
resource is being considered. The baseline case has only one intended recipient AT. The persistent
assignment avoids repeated control overhead, while sharing provides possibly better bandwidth
utilization. The format, structure, and control overhead associated with the design is under study.
Multi-user packets (MUPs) are also being considered. MUPs contain messages intended for
multiple ATs, thus extending the concept of a shared assignment. The packet contains additional
header that indicates the intended AT IDs and the segments the bits for different ATs, thus
providing coding gain for the larger packet sizes. The structure, ACK mechanism, control overhead,
and power efficiency are being studied.
1.3.8 Group Resource Allocation
The HKLLMNQRSUZ-FDD system supports low-rate delay-sensitive applications like VoIP
efficiently, by grouping such users and assigning the group a set of shared time-frequency
resources, and using bitmap signaling to allocate resources in each application frame. Such
grouping facilitates the following two forms of statistical multiplexing.
- 61-
Statistical multiplexing is achieved among the group members. When the AN has
determined a DTX state for a particular AT in a particular time period, the time-frequency
resources can be assigned to another user.
Statistical multiplexing is achieved between initial and subsequent HARQ transmissions.
Once an AT acknowledges its VoIP packet, the time-frequency resources are available for
other users.
The unused portion of the shared time-frequency resource can be temporarily assigned to other
(e.g. best effort) users.
To establish a group, the AN first assigns the group a unique identifier, namely a GroupID. Then,
the AN assigns ATs to the group, using the unique identifiers of the ATs and the unique identifier
for the group. For example, the AN can assign AT4, AT7, AT13, and AT29 to GroupID0.
As part of assigning ATs to a group, each AT is assigned a unique location within the group. For
example, AT4 can be assigned the 0th position, AT7 can be assigned the 1st position, AT13 can be
assigned the 2nd position, and AT29 can be assigned the 3rd position. The position can be thought
of as a unique identifier, which is valid only within the group.
The GroupID is used to control the entire group at once. For example, the AN can use the GroupID
to assign or change the set of shared time-frequency resources which the group uses.
Once a group of ATs is established, the AN assigns the group a set of shared time-frequency
resources and an ordering pattern indicative of the order in which the resources are allocated (if
only one ordering is allowed, the ordering pattern can be omitted). In the time domain, the set of
shared resources will be a group of VoIP frames comprising a VoIP interlace pattern. In the
frequency domain, the shared resources will typically be a set of DRCHs, although a set of BRCHs
could also be used. Figure 40 shows one possible time domain assignment (cyclic prefix of 6.51
us).
Vocoder Frame
(20 msec)
Superframe
(22.94 msec)
= VoIP Frame
= Preamble
VoIP
0 0 0 0
Interlace
VoIP 0 1 2 3
Interlace
Offset
Figure 40. Example Set of Time Domain Resources
- 62-
Referring to Figure 40, a VoIP frame is defined as two contiguous frames, which increases the
trunking gain, and the set of shared time domain resources is one VoIP interlace. Each AT is
assigned a VoIP interlace offset indicating in which VoIP frame of the interlace the transmission of
the first subpacket will occur. In the example of Figure 40, 1/4 of ATs in the group are assigned to
each of the four interlace offsets for their transmissions of the first subpackets. As a study item, it
may be advantageous to have only three interlace offsets in order to better align the time between
successive first transmissions to the vocoder frame duration.
Table 8 provides the list of fields that are needed for group setup messages for VoIP, which will be
managed through upper layer signaling carried on F-DCH.
Field Description
MAC_Index Unique identifier of the AT
Group_ID Unique identifier for the group
Block_Size The fundamental block size (e.g. 1 DRCH by 1 Frame)
Num_Blocks Number of blocks assigned to this group
First_Block Address of the first block in the assignment
Ordering_Pattern One of a few choices indicating the order in which the
blocks are to be distributed
F_Mod_Coding Coding and modulation for full rate frames
H_Mod_Coding Coding and modulation for half rate frames
Q_Mod_Coding Coding and modulation for quarter rate frames
E_Mod_Coding Coding and modulation for eighth rate frames
Interlace_Structure The pattern and structure of the VoIP interlace
Bitmap1_Length Length of the first bitmap
Bitmap2_Length Length of the second bitmap (if used)
Bitmap_Channel Time frequency resources for the bitmap itself
AT_Index The bitmap position assigned to the AT
Interlace_Offset Offset assigned to the AT indicative of its first
transmission
Table 8: Group Setup Message Fields
Once a group of ATs is established and assigned a set of shared time-frequency resources
persistently, bitmap signaling is used to assign resources to individual ATs in each VoIP frame. The
bitmap signaling is used by the AN to assign resources and by the ATs to determine their exact
resources within the set of shared time-frequency resources, and it is used for first and subsequent
- 63-
retransmissions. Persistent resources can be assigned for the first transmission of a VoIP packet
(under study - TBD).
The bitmap is jointly encoded in one packet for performance efficiency, and is sent over the span of
the VoIP interlace, unless it is more efficient in terms of overall VoIP capacity to fit in one F-SCCH
message. The time frequency resource for the bitmap is indicated by the Bitmap_Channel in the
Group Setup message.
A first bitmap is used to indicate which ATs are being served in each VoIP frame, where each AT
corresponds to a location in the bitmap. A second bitmap may be used to indicate number of
assigned resources and/or the packet format. Each AT determines its allocation based on the
allocations for all ATs with a smaller bitmap position in the first bitmap. An example of group
resource assignment is shown in Figure 41Error! Reference source not found.Error! Reference
source not found.Error! Reference source not found.Error! Reference source not found.Error!
Reference source not found.Error! Reference source not found.Error! Reference source not
found.Error! Reference source not found.. In this example, 24 ATs in a group are assigned to a
set of shared resources in one VoIP frame consisting of 8 DRCH resources (each DRCH is 16 tones
distributed in the frequency domain × 8 symbols) in each of the two adjacent frames.
The first bitmap is used to indicate active ATs. The bitmap locations correspond to the AT
positions. For example, the AT assigned the 0th group position determines its assignment based on
the 0th position in the first bitmap. Each AT with a „1‟ in the first bitmap is active. The AT with the
first „1‟ is assigned the first M blocks, the AT with the second „1‟ is assigned the second N blocks,
etc, where M and N are the same if there is only the first bitmap, and M and N may be different if
there are two bitmaps. The user with the first „1‟ in the first bitmap corresponds to the first
position in the second bitmap, the user with the second „1‟ in the first bitmap corresponds to the
second position in the second bitmap, etc. In this illustrative example, a „0‟ in the second bitmap
corresponds to an assignment of one block, while a „1‟ in the second bitmap corresponds to an
assignment of two blocks. The first active wireless terminal, AT0, is assigned one resource, and
since it‟s the first AT allocated, it is allocated block 0. The second active AT, AT2, is assigned one
block. AT2 must sum the number of resources allocated to ATs with a smaller position in the first
bitmap. In this case, AT2 must determine that one resource was previously assigned. Therefore,
AT2 is assigned block 1. The third active AT, AT4, is assigned one block. AT4 must sum the number
of resources allocated to ATs with a smaller position in the first bitmap. In this case, AT4 must
determine that 2 resources were previously assigned (1 for AT0 and 1 for AT2). Therefore, AT4 is
assigned block 2. This process is repeated for all ATs.
- 64-
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Bitmap 1 indicates
1 0 1 0 1 1 1 0 1 1 1 1 0 0 1 0 1 0 0 1 0 0 1 1 active users
Bitmap 2, if used, indicates
0
number of resources allocated to
0 0 0 1 0 0 0 0 0 0 1 0 0
each active user (0=1 resource,
1 = 2 resources)
24 ATs assigned to
this group, with
locations 0-23
Ordering Pattern
Set of shared
resources is 8 AT23 AT22
DRCH by 2
AT16 AT19
frames. Each
AT16 AT114
user determines
AT10 AT11
its allocation
DRCH Index
based on the AT9 AT8
allocations for all AT5 AT6
users with a AT5 AT4
smaller bitmap AT0 AT2
position.
Time (frames)
Figure 41: Example of Group Resource Allocation Signaling
The ATs are assigned an ACK position based on their position assignment in the first bitmap. For
example, the first N/2 ATs in the first bitmap will be assigned to transmit their ACK in the first
ACK position, while the second N/2 ATs in the first bitmap will be assigned to transmit their ACK
in the second ACK position. Similarly, an even/odd structure could be used, whereby ATs with an
odd position assignment in the first bitmap will be assigned to transmit their ACK in the first ACK
position, while ATs with an even position assignment in the first bitmap will be assigned to
transmit their ACK in the second ACK position.
1.3.9 Reverse Link Power Control (PC)
The AT transmit power is controlled in a HKLLMNQRSUZ-FDD system. The PC mechanisms
include open loop adjustments of the reference power level, a closed loop PC of the RL control
channels and RL traffic reference power level, and a delta-based PC for traffic channel. The open
loop and fast closed loop PC set the transmit power levels on the RL pilot channel that is
transmitted periodically. Other CDMA control channel power levels are set at an offset relative to
the pilot channel, and are adjusted based on the control segment rise over thermal (RoT) and PQI
information from their target sectors. The RoT value of each active sector member is initially
communicated to the AT at the time of active set addition, and can be further updated through
upper layer messages. The traffic channel power level is set at an offset relative to the pilot
channel power level, and is adjusted based on interference indications received from neighboring
sectors. The R-ACKCH and traffic channel power control also make use of the interference over
- 65-
thermal (IoT) information from FL serving sector and RL serving sector, which are broadcast over
superframe preamble. The fast closed loop PC is done at the rate of 180Hz or lower.
1.3.9.1 Open Loop Power Adjustments
The open loop power adjustments are based on the variations in the total received power, and
ensures timely response to large variations in the shadowing component of the fade. The AT
measures the total received power during the superframe preamble of the RL serving sector, and
compares it to the total received power during the previous superframe preamble of the RL serving
sector. Then, in response to a step change of P in the total received power, it adjusts the transmit
power of the RL pilot channel by -P over the open loop transition time, TOLT, i.e., using the unit
step response function min t ,1 . The open loop transition time, TOLT, should be small enough
TOLT
to allow for quick response to large variations in the shadowing, yet large enough to prevent the
open loop power adjustments from interfering with the closed loop power control, under normal
conditions.
1.3.9.2 Reverse Link Control Channel Power Control
The procedure varies depending on the channel, and is described separately for all the relevant
channels below.
1.3.9.2.1 R-PICH
AP makes use of pilot signals that are periodically transmitted on the RL CDMA control segment as
a reference for fast closed loop PC. The use of periodic control channel transmissions as reference
power level for traffic channel is necessary to control the inter-carrier-interference and to ensure
appropriate power level for minimum RL date rate.
PC commands are sent to every AT by its RL serving sector. AT uses these PC commands to adjust
its transmit power on R-PICH. Then, the AT uses the transmit PSD of R-PICH as a reference level
for all other control channels and for the traffic channel. R-PICH transmissions to other sectors (on
other control subsegments) are at the same power as the R-PICH transmission to RL serving
sector.
1.3.9.2.2 R-CQICH
1.3.9.2.2.1 To Reverse Link Serving Sector
For R-CQICH transmissions to the RL serving sector, the AT sets the transmit PSD by adding the
RoT value and an additional boost, which is specified by the RL serving sector, to the R-PICH
transmit PSD:
PR CQICH , RLSS PR PICH RoTRLSS CQIPilotBo ost
where CQIPilotBoost is a fixed boost specified in the active set assignment message, and all the
terms of given in units of dB.
- 66-
1.3.9.2.2.2 To Forward Link Serving Sector
In non-disjoint link scenario, for R-CQICH transmission to the FL serving sector, the AT uses the
same rule as the one for RL serving sector:
PR CQICH , FLSS , Nondisjo int PR PICH RoTFLSS CQIPilotBo ost .
The serving sector should be able to update the value of this boost parameter if necessary. For
temporary adjustment of the R-CQICH performance (between update messages), the FL serving
sector can also run an outer loop on the R-PICH set-point based on the R-CQICH erasure rate. To
enable accurate measurement of the R-CQICH erasure rate, in addition to the CQIReportPhase
and CQIReportPeriod parameters (which are assigned for complexity purposes), the FL serving
sector also assigns a MandatoryCQIReportPeriod, which is a multiple of CQIReportPeriod.
To ensure robust FL performance in disjoint link scenarios, for R-CQICH transmissions to the FL
serving sector in the case of disjoint link scenario, AT sets the transmit PSD based on the transmit
PSD of the R-PICH, by taking into account the RoT value of the FL serving sector and the PQI
feedback from the FL serving sector, to achieve a certain received signal to interference ratio,
specified by the FL serving sector:
PR CQICH , FLSS , Disjoint PR PICH RoTFLSS pCoTFLSS CQITargetC toI ,
where pCoT denotes the filtered pilot carrier over thermal information reported to the AT by the
corresponding sector on F-PQICH.
1.3.9.2.2.3 To Desired Forward Link Serving Sector
To ensure the reliability of the FL handoff request, for R-CQICH transmissions to the desired FL
serving sector, AT sets the transmit PSD based on the transmit PSD of the R-PICH, by taking into
account the RoT value of the target sector and the PQI feedback from the target sector, to achieve a
certain received signal to interference ratio, specified by the target sector:
PR CQICH , DFLSS PR PICH RoTDFLSS pCoTDFLSS HandoffCQI CtoI .
To reduce the erasure rate and increase the reliability of the FL handoff requests, AN may specify a
higher required signal to interference ratio for a handoff CQI than a normal CQI.
1.3.9.2.2.4 To Other Sectors
As mentioned earlier, the AT may also send R-CQICH to other sectors in the active set, for
purposes of power controlling FL control channels. For these R-CQICH transmissions, AT applies a
fixed boost relative to the R-CQICH transmissions to the RL serving sector.
1.3.9.2.3 R-SFCH and R-BFCH
For R-SFCH, and R-BFCH transmissions to the FL serving sector, AT applies a fixed boost relative
to the R-CQICH transmissions to the FL serving sector.
- 67-
1.3.9.2.4 R-REQCH
1.3.9.2.4.1 To Reverse Link Serving Sector
For R-REQCH transmissions to the RL serving sector, AT sets the transmit PSD based on the
transmit PSD of the R-PICH, by taking into account the RoT value of the RL serving sector and the
PQI feedback from the RL serving sector, to achieve a certain received signal to interference ratio,
specified by the RL serving sector:
PR REQCH ,RLSS PR PICH RoTRLSS pCoTRLSS REQTargetC toI .
1.3.9.2.4.2 To Desired Reverse Link Serving Sector
For R-REQCH transmissions to the desired RL serving sector, AT sets the transmit PSD based on
the transmit PSD of the R-PICH, by taking into account the RoT value of the target sector and the
PQI feedback from the target sector, to achieve a certain received signal to interference ratio,
specified by the target sector:
PR REQCH ,DRLSS PR PICH RoTDRLSS pCoTDRLSS HandoffREQ CtoI .
To reduce the erasure rate and increase the reliability of the RL handoff requests, AN may specify a
higher required signal to interference ratio for a handoff REQ than a normal REQ.
1.3.9.2.5 R-ACKCH
For R-ACKCH transmissions to the FL serving sector, AT sets the transmit PSD based on the
transmit PSD of the R-PICH, by taking into account the interference over thermal ratio (IoT) of the
FL serving sector, which is broadcast during the superframe preamble, and the PQI feedback from
the FL serving sector, to target a certain received signal to interference ratio, specified by the FL
serving sector:
PR ACKCH PR PICH IoTFLSS pCoTFLSS ACKTargetC toI ACKBoost PACK NACK ,
where ACKBoost is an additional boost based on the ACK to NACK error estimate, and is applied if
AT can reliably estimate the probability of the ACK to NACK error event.
1.3.9.3 Reverse Link OFDMA Traffic Channel Power Control
While traffic channel transmissions from different terminals occupy different dimensions in time
and frequency, it is not desirable to have a large difference in received power across subcarriers as
this increases the receiver dynamic range requirements and also causes loss of orthogonality with
time/frequency errors. In a multi-sector layout, high inter-sector interference can also drastically
reduce the network capacity.
Since HKLLMNQRSUZ FDD uses orthogonal multiple access for traffic channels, the serving sector
lacks information regarding the inter-sector interference caused by RL traffic originated from this
sector. Hence, it is desirable to have an interference control algorithm implemented in an AT where
the interference information could be made readily available from other sectors. In the proposed
system, a load indication is broadcasted every superframe over the F-OSICH from each sector when
the average interference over thermal level exceeds a target threshold. The load indicator takes on
- 68-
one of three values (0, 1 and 2) to control the interfering AT‟s power level. The coverage of F-OSICH
is the same as the acquisition pilots, which penetrate far into neighboring sectors. In addition, the
load indicator can also be transmitted over the FL physical channel F-FOSICH. This channel is
transmitted at a faster rate (once every PHY Frame), but has more limited coverage. The purpose of
this segment is to reach nearby ATs that can potentially cause high levels of interference.
The amount of inter-sector interference per-subcarrier caused by a given AT is determined by the
transmit power level used by that AT and the location of the AT relative to the neighbor sectors. For
the traffic channels, power control may be performed such that each AT is allowed to transmit at a
power level that is as high as possible while keeping intra-sector and inter-sector interference to
within acceptable levels. An AT located closer to its serving sector may be allowed to transmit at a
higher PSD level since this AT will likely cause less interference to neighbor sectors. Conversely, an
AT located farther away from its serving sector and toward a sector edge may be restricted transmit
at a lower power level since this AT may cause more interference to neighbor sectors. Controlling
transmit power in this manner can potentially reduce the total interference observed by each sector
while allowing “qualified” ATs to achieve higher SNRs and thus higher data rates (see Figure 42).
The transmit Power Spectral Density (PSD) (defined as the transmit power per assigned subcarrier)
for a traffic channel for a given AT may be expressed as:
Pdch (n ) Pref (n ) P( n ) ,
where Pdch ( n ) is the transmit PSD for the traffic channel for update interval n;
Pref (n ) is a reference PSD level for update interval n; and
P(n) is a transmit PSD delta for update interval n.
The PSD levels Pdch ( n ) and Pref ( n ) and the transmit power delta P(n) are given in units of
decibels (dB/Hz). The power control algorithm described in this document will be called Delta-
based power control because of the transmit power delta P(n) . The reference PSD level provided
by the RL pilot channel in this design.
- 69-
Figure 42: Reverse Link Traffic Channel Power Control
The transmit PSD for the traffic channel is set based on the amount of inter-sector interference the
AT may be causing to other ATs in neighbor sectors, the amount of intra-sector interference the AT
may be causing to other ATs in the same sector, and the maximum power level allowed for the AT.
In order to keep the intra-sector interference at acceptable levels, the transmit PSD delta, P(n) ,
is constrained to be within a range as follows: P( n ) [ Pmin , Pmax ] , where Pm in / Pm ax is the
minimum/maximum transmit PSD delta allowable for a traffic channel. The Pm in and
Pm ax values are regularly adjusted based on the PQI feedback on F-PQICH and the IoT value of the
RL serving sector, to make the minimum and maximum traffic signal to interference ratio
independent from the R-PICH set point and IoT variations at the RL serving sector.
Each sector can estimate the average amount of interference relative to thermal noise power
(referred to as IOT) experienced by that sector from terminals in other sectors, where the thermal
noise level is measured during the RL Silence Interval. Each sector broadcasts an indication of its
interference measurements for use by ATs in other sectors. For simplicity, the following description
assumes the use of a single load indicator bit to provide interference information. Each sector may
set its other sector indication (OSI) as follows:
'1' or '2', if IOTmeas,m (n) IOTtarget , and
OSIm (n)
'0' ,
if IOTmeas,m (n) IOTtarget ,
- 70-
where IOT m eas,m ( n ) is the measured IOT for sector m in time interval n; and IOT target is the desired
operating point for the sector. OSI value of „2‟ is used to indicate excessive IOT level. Here IOT
refers to the interference-over-thermal, which is a ratio of the total interference power observed by
the sector to the thermal noise power.
Each AT can estimate the channel gain (or propagation path gain) for each sector that may be
interfered by the AT. For a user, the channel gain ratio between the serving sector and a neighbor
sector may be viewed as a “relative distance” that is indicative of the distance to a neighbor sector
relative to the distance to the serving sector.
In one setup, each AT monitors the OSI broadcast by neighbor sectors and only responds to the
OSI of the strongest neighbor sector, which has the smallest channel gain ratio. If the OSI from
that sector is set to „1‟ or „2‟ (due to the sector observing higher than nominal inter-sector
interference), then the AT adjusts its delta downward. Conversely, if the OSI is set to „0‟, then the
AT adjusts its delta upward.
The load indicator bit thus determines the direction in which to adjust the transmit power. The
amount of transmit power adjustment for each AT may be dependent on the current transmit
power level (or the current transmit power delta) of the AT and the channel gain ratio for the
strongest neighbor sector.
The AT sends the transmit PSD delta and the maximum number of subcarriers that the AT can
support at the current transmit PSD delta via in-band signaling. This information is used by the
AN for making RL assignments. Thus, an AT having a low delta may be assigned a large number of
subcarriers so that it can use all its transmit power to achieve a higher data rate.
This mechanism has the following advantages: explicit interference control leads to tight
interference tail distribution, while Delta-based power control naturally shapes the PSD of users
that cause high and low interferences, while centralized scheduling achieves the desired
fairness/capacity tradeoff.
1.3.9.3.1 RDCH Gain Determination
The OSIMonitorSet provides information on the other sector interference. After each OSIMonitorSet
update, the AT creates an OSI vector whose ith element corresponds to the most recent OSIValue
from the ith sector. The most recent OSIValue can be a value received over the F-OSICH or
F-FOSICH of the sector. In addition, the AT creates a ChanDiff vector whose ith element is
indicative of the difference in the channel strengths from that AT to its serving sector and the ith
sector. For sectors from which AT receives a reliable PQI information on F-PQICH, the ChanDiff
value in dB is simply the difference between the PQI values received from the RL serving sector and
that sector. For all other sectors, the ChanDiff is computed as:
RxPowerRL,SS TransmitPoweri
ChanDiffi
TransmitPowerRL,SS RxPoweri
where RxPowerRLSS and RxPoweri are the average received power (across antenna) of the F-ACQCH
of the RL serving sector (RLSS), and the average received power (across antenna) of the F-ACQCH
of the ith sector. TransmitPowerRLSS and TransmitPoweri are the average transmit power of the F-
ACQCH of the RLSS, and the average transmit power of the F-ACQCH of the ith sector.
- 71-
In absence of OSI information, maximum transmit power is used. Else the AT computes a Decision
Threshold vector as follows:
max{UpDecisionThresholdMin, (1 a)bi } if OSI i 0
DecisionThreshold i max{DownDecisionThresholdMin, a(1 bi )} if OSI i 1
OSI i 2
1 if
where UpDecisionThresholdMin and DownDecisionThresholdMin are configuration attributes of the
protocol. Variables a and bi are determined as follows:
min{RDCHGain, RDCHGainMa } RDCHGainMi
x n
a , and
RDCHGainMa RDCHGainMi
x n
min{ ChanDiff i , ChanDiffMa x} ChanDiffMi n
bi ,
ChanDiffMa x ChanDiffMi n
where ChanDiffMax and ChanDiffMin are configuration attributes of the protocol. The AT computes
a Decision vector whose ith element is given by:
UpDecisionValue if xi DecisionThreshold i and OSI i 0
Decisioni DownDecisionValue if xi DecisionThreshold i and OSI i 1 or 2
0 otherwise
where 0 xi 1 is a uniform random variable and UpDecisionValue and DownDecisionValue are
configuration attributes of the protocol. The AT then computes a weighted decision, Dw, by:
OSIMonitorSetSize
1
i 1 ChanDiffi
Decisioni
Dw OSIMonitorSetSize
1
ChanDiff
i 1 i
The AT increases RDCHGain by a preset amount if Dw is greater than this threshold, and decreases
RDCHGain by another preset amount if Dw is less than or equal to the negative of the threshold.
The RDCHGain always lies between two preset values to satisfy ICI margin requirements.
To reduce the packet error rate, the RL serving sector may assign separate boost values for
different H-ARQ attempts, in which case, AT applies the corresponding boost value for each H-ARQ
transmission, after the RDCHGain adjustment explained above.
1.3.9.4 Reverse Link CDMA Traffic PC
The RL CDMA traffic usually carries low rate delay sensitive data and is sent on R-CDCH. The ratio
of R-CDCH to the R-AuxPICH is fixed based on the packet format, and is controlled by the outer
loop. This also determines the power offset between the R-AuxPICH) and the R-PICH. The T/P ratio
- 72-
can be updated by the AT based on H-ARQ termination (F-ACKCH statistics). An option to DTX fast
PC commands when CDMA traffic is being transmitted is under study (TBD).
Loading control is done at the AN based on admission control and the use of slow RAB bits for
tighter control of CDMA traffic (TBD). As described before, a fast closed loop power control is used
to set the transmit power level on the RL control channels that are transmitted periodically.
1.3.10 Fractional Frequency Reuse
Although HKLLMNQRSUZ FDD is robust to interference and can be deployed with universal
frequency reuse, frequency planning could be used to enhance coverage and QoS. Frequency reuse
is often used in interference limited systems to improve channel C/I, hence improving link
reliability at sector edge. The resulting channel quality improvement, however, comes at a cost of
bandwidth reduction, which is not necessarily a good capacity tradeoff. For example, a system with
1/3 frequency reuse needs to improve an AT‟s spectral efficiency by 200% to achieves the same
throughput as a 1/1 reuse system. According to the AWGN capacity formula, a 200% spectral
efficiency gain requires at least 5 dB gain in C/I in the linear (low SNR) regime, and each bps/Hz
improvement in the nonlinear (high SNR) regime asymptotically requires 3 dB gain in C/I.
Fractional frequency reuse enables ATs with different channel condition to enjoy different reuse
factor within the same AN. A small fraction of bandwidth and power resources can be assigned to
the ATs that benefit from interference reduction achieved by reuse. Examples are AT at the cell
edge, on FL as well as RL, and ATs with MIMO capabilities which can advantageously transform
high C/I into a gain in terms of spectral efficiency. Details regarding the support of FL and RL FFR
are under study.
1.3.11 Subband Scheduling
When scheduling multiple ATs in frequency selective channels, system capacity can be increased
by scheduling each AT in a preferred subband based on its current channel frequency response.
Subband scheduling improves fairness through SNR gains to weak users, and enables multi-user
diversity gains for latency sensitive users. The design supports two hopping modes: a diversity
mode with global hopping across the band, and a subband mode with localized hopping, with the
subband size being configurable (details TBD).
1.3.11.1 Local Hopping and Channel Trees
In general a larger number of subbands potentially increase the ability to get larger multi-user
gains. However, the subband size should provide enough frequency diversity to prevent
performance degradation for fast moving ATs, and potential increase in feedback overhead. Another
implication of having too narrow subbands is a loss in trunking efficiency since there would be
fewer candidate ATs to be scheduled per subband. As an example, a subband size around
1.25 MHz (128 contiguous tones) is used for 256pt FFT and above and around 640kHz (64
contiguous tones) for 128pt FFT. The following description refers to the case of 256pt FFT and
above while extension to the 128pt FFT case is straightforward. When an AT scheduled within a
subband with a bandwidth assignment less than the entire subband, its assignment will hop
„locally‟ across this subband in order to maximize channel and interference diversity.
- 73-
Based on the above example, channel trees are defined with local hopping so that all parent nodes
that host 8 base nodes (hence 128 subcarriers) will map to fixed subbands of 128 contiguous
subcarriers. However, channels corresponding to the base nodes underneath this parent node will
be hopping across the subband (see Figure 43) to ensure interference diversity.
Figure 43: Subband Scheduling
1.3.11.2 Subband Feedback
In order to support subband scheduling, the AT needs to provide feedback about the FL channel
properties relative to different subbands. The amount of feedback should balance gains in FL
performance due to subband scheduling versus the RL overhead caused by feedback channels. The
R-SFCH includes a subband index which is used by AT to indicate the preferred subband. Based
on this indication, the AN may schedule the AT over this subband with rate determined from the R-
CQICH. An extended version of the R-SFCH includes a subband channel quality indicator in
addition to the subband index, i.e. a flexible reporting size is allowed. This channel allow for
variable reporting interval. The average reporting interval is defined by the AN.
The RL CDMA broadband pilot enables FL subband channel quality assessment at the AN. For RL
subband scheduling, the rate prediction and scheduling is done by the AN. It is possible to use RL
CDMA control channel R-PICH or the supplemental RL broadband CDMA pilot for RL subband
channel quality assessment at the AN.
1.3.12 Quasi Orthogonal Reverse Link
While OFDMA benefits from the elimination of intra-sector interference, it has the disadvantage of
becoming dimension limited as the number of receive antennae at the AN increases. As the number
of receive antennae increases, a non-orthogonal scheme like CDMA can operate in the interference
limited regime where the SNR per modulation symbol is low, and the system capacity scales
linearly with SNR – thus the capacity increases linearly with the number of receive antennae. In an
orthogonal RL, the total load is limited by the system bandwidth, while spectral efficiency per
dimension scales logarithmically with SNR, thus the system capacity scales logarithmically with
the number of antennae.
- 74-
This fundamental limitation of orthogonal multiple access is mitigated by using a quasi-orthogonal
multiplexing scheme on the RL by assigning the same bandwidth to multiple ATs in the same
sector. Spatial processing is used to demodulate these ATs leading to a design that reaps the
benefits of an orthogonal design when the number of receive antennae is small, and offers
improved capacity scaling with the number of antennae.
1.3.12.1 Random Hopping
Intra-sector interference diversity is obtained through random hopping. AN assigns each AT a set
of time-frequency blocks that hop in frequency over time. When such hopping sequences assigned
to multiple ATs overlap, every AT will overlap with a set of ATs on every time-frequency block. The
sets of ATs will be different for different blocks, hence providing co-channel interference diversity
which is advantageously used by H-ARQ to terminate packet transmission at an appropriate rate.
Multiple antennae are used to suppress intra-sector interference using a space-frequency MMSE
receiver. The concept of random hopping should be implemented to support different values of the
multiplexing factor (e.g. the number of ATs assigned the same time-frequency blocks). It is also
important to ensure co-existence of quasi-orthogonal with orthogonal assignments which may be
needed to support high QoS requirements.
1.3.12.2 Multiplexing Factor Control Through Scheduling
The channel assignment structure is defined by a channel tree, each base node of which maps to a
time-frequency block. The mapping from base nodes to time-frequency blocks is randomized in
time, thereby resulting in frequency hopping for each assignment. In the quasi-orthogonal mode,
one possibility is to use a channel tree that contains Q identical sub-trees as illustrated in Figure
44 such that base nodes of each sub-tree are randomly mapped to the same set of time-frequency
blocks. Within each sub-tree, the base nodes map to disjoint resources.
Figure 44: Multiple Channel Trees to Support Quasi-orthogonal Operation
For orthogonal operation, ATs are scheduled on a single sub-tree. In quasi-orthogonal mode, an
integer multiplexing factor Q can be achieved by loading exactly Q sub-trees. A fractional value of
Q can be achieved through a symmetric partial loading of Q sub-trees, where . is the integer
ceiling operation. Alternatively, fractional Q can be achieved through asymmetric partial loading so
that the RL scheduler starts assigning resources to an AT on a new sub-tree, in addition to the
existing fully loaded sub-trees.
- 75-
1.3.12.3 Orthogonal Pilot Multiplexing
In order to optimize demodulation performance for the quasi-orthogonal scheme, pilots should be
designed so as to enable accurate estimation of Q channels corresponding to the ATs multiplexed
over a block. These channel estimates will be used to set the parameters of a linear receiver (such
as MMSE) or a non-linear (e.g. successive cancellation) receiver at the AN. In OFDMA, RL traffic
resources can be assigned in units of time-frequency blocks with local (dedicated) pilots placed in
every block for channel and interference estimation, as orthogonal pilots to improve channel
estimation. This consists of defining a few contiguous clusters of pilots with cluster location
optimized to minimize channel estimation error in orthogonal mode and to multiplex pilot
dimensions of different ATs over these clusters, by using some orthogonal codes. Different
sequences of orthogonal codes will be implicitly assigned to different ATs if every channel tree
described in the previous section is associated with an orthogonal sequence. An example of time-
frequency block and pilot design that supports quasi-orthogonal operation with Q=3 is shown in
Figure 45. Here the pilot symbols are arranged in six clusters with three “strips” in frequency and
two “strips” in time. In quasi-orthogonal mode, these clusters can be used to orthogonally
multiplex pilots of different ATs. This multiplexing can be achieved by assigning e.g. different
vectors of a 3x3 orthogonal basis (such as three-dimensional DFT basis) to different channel sub-
trees.
Figure 45: Pilot and Data Symbol Placement in a Time-Frequency Traffic Block
This pilot structure can be also used to facilitate decoding of RL transmission by different sectors
of the serving cell. This mode of operation known as softer handoff helps to improve system
coverage and link budget for ATs located close to the sector boundaries. In orthogonal mode,
different sectors of the same cell will use the same pilot/data symbol multiplexing with orthogonal
pilot multiplexing for different sectors of the same cell. Such an orthogonal multiplexing allows any
sector to accurately extract channel state corresponding to RL transmission within this sector as
well as RL transmissions taking place in the adjacent sectors. Hence, traffic demodulation within a
given sector can be assisted by a neighbor sector that will use appropriate receiver architecture to
separate data symbols transmitted within the two sectors. Performance of softer handoff can be
enhanced substantially when multiple receive antennae are used in every sector AN to enable
spatial separation of data symbols transmitted in different sectors.
The orthogonal pilot patterns can be further optimized based on evaluation results (TBD). The use
of resource orthogonal pilots is under study (TBD).
- 76-
1.3.13 Layered Superposed OFDMA (LS-OFDMA)
As an extension to Quasi Orthogonal RL, overloading of OFDMA resources by letting users overlap
in time/frequency is allowed, and is called LS-OFDMA. The key idea is to “layer” users according to
achievable spectral efficiency. Overloading is possible per spatial dimension and therefore can be
used with or without multiple antenna receivers. Users with similar spectral efficiency are
orthogonalized within the same layer and intra-layer fairness as in a conventional OFDMA system
with scheduling. Users in different layers interfere with each other and occupy the overlapping
bandwidth. Data of different users/layers are recovered due to interference cancellation or joint
decoding. Packet format and bandwidth to all users are explicitly allocated by the AN.
The layers hop independently of each other. Independent hopping of layers is achieved by
scheduling ATs of different layers on different sub-trees of the channel tree that same mapped to
the same set of resources (tiles) by different hopping patterns. Note that independent hopping helps
in achieving better fairness as the interference cancellation gain from a single user decoding in one
layer is spread evenly across many users in the next layer(s). Subband hopping allows for joint
hopping of large assignments that span on or multiple subbands. This feature facilitates joint
decoding.
The first layer can be assigned higher spectral efficiencies and possibly target earlier H-ARQ
terminations. The decoder attempts to first decode all first layer users, cancel those that succeed
and then proceed to the next layer. Channel estimation is performed for up to first three layers of
users based on orthogonal pilot resources. Pilot cancellation is performed based on these
estimates, and channel estimation is done for the other layers after pilot interference cancellation.
The number of non-orthogonal layers will be dynamically determined by AP. More details are TBD.
1.3.14 Handoff Scenarios
In this section, some techniques used to handle handoff between AN sectors in the HKLLMNQRSUZ
system are described. HKLLMNQRSUZ supports fast handoff to minimize the handoff impact on
latency-sensitive traffic and the response time of the system to rapid variations in the path loss
and shadowing components of the channels of serving and interfering sectors at vehicular speeds,
especially with frequency reuse of one, and to get fast fading gains and improve diversity at
pedestrian speeds. Average handoff delay can be as low as 8 ms.
An AT can be served by different sectors on the FL and RL to achieve best cell site selection gains
in situations where the best serving sectors for FL and RL are different. This is achieved at
relatively low signaling overhead, which is especially more important when fast handoff is used to
obtain fast fading gains at pedestrian speeds, or when the handoff rate is high due to high
vehicular speeds.
The sector from which the AT received the last FL Assignment Message (FLAM) is referred to as the
FL serving sector, and the sector from which the AT received the last RL Assignment Message
(RLAM) is referred to as the RL serving sector. The active set of an AT is the set of sectors that have
allocated MAC IDs and dedicated control resources to the AT. The AT monitors the F-SCCH from
the FL and RL serving sectors, as well as the desired FL and RL serving sectors. The term handoff
is used to refer to a change in the AT‟s FL or RL serving sector, which occurs when the AT receives
a link assignment message from its desired serving sector and the desired serving sector is
- 77-
different from the current serving sector. The AT can perform a handoff to any member of the
active set. The FL and RL serving sectors of an AT can also be switched within a cell by the AN
through special upper layer messages.
On a handoff, a new sector in the active set becomes the serving sector. One objective of fast
switching is to not introduce any packet loss at the higher layers while allowing for uninterrupted
transmission to the AT. A second objective is to minimize the backhaul communication required
between the APs in the active set. There is a trade off between the amount of back haul
communication required and the handoff latencies that can be achieved.
In order to continue uninterrupted data transmission, the new serving sector needs to know the
forward looking RLP state for the FL, where the forward looking RLP state is defined as the data
received at the anchor AP and not yet transmitted, and the data that needs to be retransmitted
based on ReceiverStatus messages from the AT.
1.3.14.1 Active Set Management
In this subsection, the active set management protocol is briefly described.
The AT performs FL SINR measurements on FL pilots suggested by the AN, and on
pilots it autonomously searches. The AT filters the measured values to remove
measurement noise and fast fading components.
Active set of an AT is determined based on FL pilot measurements.
The AN assigns dedicated control resources to the members of the active set for the AT,
assigns a MACID for the AT on each member of the active set and transmits an active
set assignment message to the AT through its FL serving sector.
The AT monitors assignment channels from the members of the active set that are
currently serving the AT on forward and reverse link, as well as the desired serving
sectors on both links.
The handoff procedures described in the following two sub-sections apply to the AT initiated
handoffs, when the different sectors are synchronous with each other.
1.3.14.2 Forward Link Handoff
The FL handoff is performed in the following steps:
Selection: FL serving sector selection is primarily based on the FL pilot strength
measurements. The AT can use the acquisition pilots (similar to the active set
management algorithm) and/or broadband CQI pilots. In selecting the FL serving
sector, AT has to also take into account the RL channel qualities, communicated
through F-PQICH, to make sure that it will not cause too much interference at other
sectors, i.e.,
pCoTMax pCoTDFLSS MaxRLPilot Difference ,
where pCoTMax is the largest pCoT among the active set members of the AT.
Indication: The AT requests a FL handoff by targeting R-CQICH to its desired FL
serving sector and setting the one-bit Desired FL Serving Sector (DFLSS) flag.
- 78-
Signaling and Detection: A successful decoding of R-CQICH carrying handoff request
at the desired FL serving sector indicates a FL handoff request.
RLP State Discovery at the Desired FL Serving Sector: Details will be considered in
Working Group 2.
Completion: Once the desired serving sector is ready to serve the AT, it sends a FLAM
to that AT on F-SCCH, by which it indicates the completion of the FL serving sector
switch or L1 handoff.
1.3.14.2.1 Softer Handoff Groups
The FL channels can be sent in softer handoff – this is intended to improve coverage at the cell
edges. The AN may advertise softer handoff groups with the following interpretation: The AN
ensures SFN transmission within the group to a handoff AT (i.e. the sectors in the handoff group
transmit the same waveform to the handoff user as serving sector, and the BTS can use various
forms of diversity, e.g. delay diversity, phase sweep diversity, etc.) and the AN ensures that no
interference will be seen by a handoff AT from group members.
The AT can make handoff decisions and request the serving group based on the combined pilot
measurements from all group members to compute the anticipated SFN channel strength and
predict interference level given, e.g., by the overall received power level less the contribution of the
group based on the pilot measurements from all group members.
The R-CQICH channel carries channel quality reports corresponding to individual sectors.
1.3.14.3 Reverse Link Handoff
On the RL, handoff is performed in the following steps:
Selection: RL serving sector selection by an AT is primarily based on the RL quality
indicator sent by every active set member of this AT. In selecting RL serving sector, AT
has to take into account the IoT value of the desired RL serving sector, to make sure
that it can meet the QoS requirements of the traffic flows it is currently running, i.e.,
PMax PR PICH pCoTDRLSS IoTDRLSS RequiredTotalPowerOverInterfer encePSD ,
where PMax is the maximum transmit power of the terminal, and
RequiredTotalPowerOverInterferencePSD is set based on the QoS requirement of the
traffic flows of the terminal. Also, to avoid causing extra interference at other sectors,
AT should not select a very weak sector as its RL serving sector, i.e.,
pCoTMax pCoTDRLSS MaxRLPilotDifference .
Indication: AT requests RL handoff by targeting R-REQCH to its desired RL serving
sector.
Signaling and Detection: A successful decoding of R-REQCH targeted to AP which is
not the current RL serving sector indicates a RL handoff request.
Completion: Upon successful detection of R-REQCH, the desired RL serving sector
completes the handoff by sending RLAM to the AT.
- 79-
2 EPILOGUE
In this document, an overview of a system designed for mobile broadband wireless access was
presented. The PHY and MAC have been designed to provide true broadband experience in the
challenging mobile wireless environment. The system supports different levels of QoS and seamless
connectivity for users across cells in a wide area environment. The PHY employs advanced
transmission techniques and features a system design that supports implementation of those
features with low system overheads. A key feature is the ability to simultaneously support
terminals employing techniques such as MIMO, beamforming, precoding, SDMA, and subband
scheduling. This enables the system to adaptively choose the appropriate transmission techniques
for different terminals based on their propagation environments.
- 80-
3 REFERENCES
[1] C30-20060626-054R1, “Framework Proposal for LBC mode of Rev C”, Miyazaki, June 2006.
[2] C.S0024-B, “cdma2000 High Rate Packet Data Air Interface Specification”, May 2006.
[3] C.S0054-A, “cdma2000 High Rate Broadcast-Multicast Packet Data Air Interface
Specification”, March 2006.
- 81-
4 APPENDIX
This section lists some details of various components of the proposal.
4.1 Precoding and SDMA Codebooks
TBD
4.2 Coding and Modulation for the R/F-DCH
The general modulation and coding schemes used in HKLLMNQRSUZ is shown in Figure 46.
Sub- CRC Channel Sequence Data
Encoder
packet 0 Insertion Interleaver Repetition Scrambler
Sub- CRC
Encoder
Channel Sequence Data Multiplexer
Input packet Packet packet 1 Insertion Interleaver Repetition Scrambler and Output Modulation
Splitting Symbol Symbols
Mapper
Sub- CRC Channel Sequence Data
Encoder
packet t-1 Insertion Interleaver Repetition Scrambler
Figure 46: Coding and Modulation Structure
4.2.1 Channel Coding
Convolutional or turbo codes are used depending on whether the packet size in less than or greater
than 128 bits. LDPC codes are under study.
4.2.1.1 Packet Splitting
Large packets are split into smaller subpackets and separately encoded. This permits different
turbo decoder units to operate on the different subpackets in parallel and thus speeds up
decoding. Furthermore, each subpacket has its own CRC in order to allow for early termination of
decoding. All subpackets have approximately the same size. The maximum possible size of a
subpacket is 8192 bits.
4.2.1.2 Convolutional Code
A rate-1/3 convolutional code is used to encode short packets in which the number of information
bits is less than or equal to 128. Such packets are primarily used in control channels. The
generator polynomial of this code is:
G(D) = [g0(D) g1(D) g2(D) ]
where g0(D)= 1 + D2 + D3 + D5 + D6 + D7 + D8, g1(D)= 1 + D + D3 + D4 + D7 + D8, and g2(D)= 1 + D +
D2 + D5 + D8, where D represents the delay operator.
4.2.1.3 Turbo Code
A rate-1/5 turbo code is used to encode packets (or subpackets) in which the number of
information bits is greater than 128. The turbo code is a parallel concatenation of two constituent
- 82-
systematic recursive convolutional codes with the turbo interleaver preceding the second recursive
convolutional encoder. The transfer function for either constituent code is:
n0 (D) n1(D)
G(D) 1
d(D) d(D)
where d(D) = 1 + D2 + D3, n0(D) = 1 + D + D3, and n1(D) = 1 + D + D2 + D3, where D represents
the delay operator. The turbo interleaver uses a combination of bit-reversal interleaving, row-
column interleaving and linear congruential interleaving. The interleaver is flexible and allows for a
wide range of packet sizes.
4.2.2 Channel Interleaving
The proposed channel interleaver design is based on a pruned bit reversal interleaver (PBRI). An m-
bit reversal interleaver operates on a sequence of length N = 2m by mapping each index to its bit
reversed value. An m-bit reversal interleaver π satisfies the following distance property: If |a-b| <
2i, then | π(a) – π(b) | ≥ 2m-i-1. This distance property is desirable since it provides for robustness
against bursty errors.
Moreover, it is easy to see that after interleaving, the first N/2 bits are those with indices divisible
by 2, the first N/4 bits have indices divisible by 4 and so on. Therefore puncturing all but the first
N/2i bits would result in transmission of bits spaced 2i apart. Thus regular puncture patterns can
be achieved when the “puncturing factor” is a power of 2. Regular puncture patterns are important
because the performance of turbo and convolutional codes under regular puncturing is better than
the performance under irregular puncturing.
A pruned bit reversal interleaver (PBRI) operates of sequences of length N when N is not a power of
2 by appending dummy symbols, using a bit-reversal interleaver and then deleting the dummy
symbols. A PBRI has the same desirable properties as the bit reversal interleaver.
The channel interleaver used with the convolutional code uses a PBRI on all the encoded bits. The
channel interleaver used with the turbo encoder splits the turbo-encoded bits into three groups as
shown below and then uses a PBRI to interleave each of these groups. The systematic bits are then
transmitted first, followed by one of the groups of non-systematic bits and then the other non-
systematic group. In addition to providing good distance and regular puncture patterns, this
channel interleaver also has the property that the code reduces to a rate-1/3 turbo code when the
transmitted bits are punctured down to that rate.
4.2.3 Incremental Redundancy by Puncturing and Repetition
Hybrid-ARQ in this air-link uses incremental redundancy. A certain fraction of the interleaved bits
are transmitted in each H-ARQ transmission. The code rate seen by the receiver goes down as the
number of H-ARQ transmissions increases. Code rates below 1/5 are achieved by repetition i.e.,
transmitting some of the encoded bits more than once. This is done only after all the encoded bits
are transmitted at least once.
- 83-
4.2.4 Data Scrambling
The encoded bits are XORed with a pseudorandom sequence prior to modulation. The
pseudorandom sequence is generated using a PN register, which is seeded with the MAC ID of the
user. MAC ID based scrambling prevents a user from accidentally decoding a packet intended for
another user.
4.2.5 Modulation Formats and Modulation Step-down
Four modulation formats (QPSK, 8PSK, 16QAM and 64QAM) are supported in the air-link. The
signal constellation for each of these modulation formats is based on the Gray mapping. The use of
multiple modulation formats and coding rates allows for a range of spectral efficiencies (also
referred to as packet formats). The system supports 15 packet formats on the FL (in SISO mode) as
well as on the RL. The number of bits in a physical layer transmission is determined by the
number of subcarriers assigned for the transmission, and the packet format chosen.
The system uses synchronous, non-adaptive H-ARQ, i.e. the channel assignments and hence
spectral efficiencies are the same for all H-ARQ transmissions of a PHY packet. Some packet
formats use lower order modulation formats for later H-ARQ retransmissions, a technique known
as “Modulation step-down.” Since the code rate is higher with a lower-order modulation format (for
the same spectral efficiency), modulation step-down helps in preventing the code rate from falling
below 1/5. The disadvantage of having a code rate lower than 1/5 is the need to repeat bits, which
is suboptimal from an information theoretic point of view. Modulation step-down thus prevents (or
at least minimizes) repetitions and thereby improves code performance.
Modulation step-down is supported on both FL and RLs, as can be seen from the following tables of
packet formats used on the FL (Table 6 - SISO mode) and RL (Table 7) respectively. Note that some
of the packet formats cannot be decoded after the first transmission because the effective code rate
at that point is greater than 1.
4.3 GCL Sequences
The Generalized Chirp Like (GCL) sequences are transmitted on TDM1 which is used for initial
acquisition and coarse frequency offset estimation (see Section 1.3.1). The k th sample of the GCL
sequence in frequency is given by
k k 1 N FFT
X (k ) exp j 2 u , k 0,1, , 1, u 1, 2, , NG 1 , N G is the smallest prime
2 NG 2
N FFT N FFT
number larger than . Taking the IFFT of the GCL sequence, and assuming NG , it can
2 2
th
be shown that the i sample in time is given by
i i 2 / u NG 1 k k 1 N FFT
x(i) exp j 2 exp j 2 u , i 0,1, , 1 .
2 NG k 0 2 NG 2
This implies that the magnitude of the time domain samples is constant independent of i . Thus
the GCL sequences have low peak to average ratio that would enable the AN to power boost TDM1.
- 84-
N FFT
Since is strictly smaller than N G , the constant modulus property of the time samples is
2
distorted. However, the distortion effect is minimal if the difference is small. The constant modulus
property in time is also distorted by the presence of guard tones. The GCL sequences also have
ideal cyclic auto-correlation and good cross-correlation properties.
4.4 Receiver Chain for R-OFDM
A brief description of the receiver chain for R-OFDM is provided for receivers employing an
advanced architecture for implementing R-OFDM.
4.4.1 Multi Dimensional Demodulator
TBD.
4.4.2 Twin Turbo Decoder
TBD.
4.4.3 Receive Chain for R-OFDM employing MIMO
TBD.
- 85-
