Uplink Power Control in LTE Relay Enhanced Cells

Document Sample
Uplink Power Control in LTE Relay Enhanced Cells Powered By Docstoc
					Uplink Power Control in LTE
Relay Enhanced Cells
Masters Thesis Presentation
Department of Communications and Networking

Student:      Aydin Karaer

Supervisor:   Prof. Jyri Hämäläinen / HUT
Instructor:   Doc. Simone Redana / NSN
   LTE Advanced
   Why Relaying?
   Relay Enhanced Cell (REC) Scenario
   LTE Uplink Power Control
   Simulation Parameters
   Power Control Optimization and System Performance
    Evaluation in REC
LTE Advanced (LTE-A)
   First set of requirements were addressed in June 2008
   Not a revolution but an evolution of LTE
   Promises to support peak data rates of 1 Gbps in downlink and 500
    Mbps in uplink
   Bandwidth scalability up to 100 MHz
   Improved user and control plane latencies
   Improved cell edge performance
Why Relaying? (1)

   Future wireless communication systems are operating
    carrier frequencies over 2 GHz
       Heavy pathloss in radio transmission
   Aggressive propagation conditions restrict the radio
    coverage especially in urban areas
   Possible solutions:
           Power increase? => Interference, decreased battery lifetime
           More base stations? => Deployment and maintenance costs,
            possibly not enough subscribers, no cell edge performance
   Relay nodes (RN) provide an attractive solution
       to satisfy tough throughput and coverage requirements for LTE-A
Why Relaying? (2)

   Result              Reasoning

     Relays can             Less eNBs and smaller OPEX
     save costs

    Relaying and            OFDMA flexible enough to fine tune e.g.
    technology fit           resource allocation
    Relay OPEX
    will be very
                            Easier to find sites, no backhaul costs, easier
         low                 installation

    Relays will be          Due to low TX power, no auxiliary equipment
Why Relaying? (3)

   Relays introduce extra delay and overhead
   Resource partitioning and interference management
    becomes important
   Deployment is challenging
   Additional set of signaling protocols are needed in case
    of Layer 2 and Layer 3 relays
   Increasing number of hops introduce more complexity
    and overhead in the system
Relay Enhanced Cell (REC) Scenario
   Simple infrastructure multi-hop scenario is considered
    with decode-and-forward relays

      Two-hop relay based deployment   Downlink received signal power in REC

   Idea is to deploy the relay nodes in the cell edges in
    order to improve the low SINR experienced by users and
    minimize the cell outage
LTE Uplink Power Control (1)

   Full frequency reuse (reuse one) is highly desirable for
    future communication systems so as to exploit the
    spectrum efficiently
   Intra-cell interference was the limiting factor in WDCMA
   LTE uplink transmission scheme SC-FDMA mitigates
         intra-cell interference
         near far effect
   However, the LTE system is sensitive to inter-cell
LTE Uplink Power Control (2)

   Standardized LTE uplink power control formula is simple
    and robust:
       P  min{ Pmax , P0  10 log10 M  L   TF ( i )  f ( i )}
   Fractional power control (FPC) utilizes a compensation
    factor for the pathloss and it is introduced to improve the
    performance of cell center users by inducing an
    acceptable inter-cell interference
   Open loop power control is considered in this work, thus
    closed loop corrections are omitted
         Used formula is given as:
      P  min{ Pmax , P0  10 log M  L)
Power Control in REC (1)

   REC requires detailed dimensioning and planning
   New cell edges introduced by RNs will lead to severe
    intra-cell and inter-cell interference
       in particular when high number of relay nodes are deployed in the
        cell with reuse one
   Power control becomes an important means in the uplink
    transmission of REC
       to mitigate the interference and increase the cell edge and system
   Approved LTE uplink power control scheme should be re-
    investigated in REC to achieve an optimal performance
       in this work, approved LTE uplink power control formula is applied
        in each relay node
   Power Control in REC (2)
   Main Simulation Parameters
    PARAMETER (Ref.1)               ASSUMPTIONS

    System Layout                   19 cells & 3 sectors/cell & 1 tier (9 RNs) of RNs/sector

    Carrier Frequency               2 GHz

    Propagation Scenario            Macro 1 (500m ISD)

    Frequency Planning              Reuse one (each eNB and RN uplink transmission interferes with each other)

    System Bandwidth                10 MHz (48 PRBs for data)

                                    eNB-UE => L  128 .1  37 .6 log 10 R (R in km)            eNB height/location = 25 m (above rooftop)
    Channel Models
                                    eNB-RN => L  124 .5  37 .6 log 10 R (R in km)            RN height/location = 5 m (below rooftop)
                                    RN-UE =>                              (R in km)            UE height/location = 1.5 m
                                                L  140 .7  36 .7 log 10 R

    Antenna Configurations                           2            Am  25 dB
                                                                                                eNB antennas per sector = 2 tx, 2 rx
                                    A( )   min 12       , Am 
                                                     3dB 
    (Pattern & Number of Ant.)                                                                  RN antennas per sector = 2 tx, 2 rx
                                                                       3dB  70 o
                                                                                            UE antennas             = 1 tx, 2 rx

    UE Transmit Power               23 dBm

    eNB Transmit Power              46 dBm

    RN Transmit Power               30 dBm

    Extra Margins                   0 dB (No shadow fading, fast fading)

    User Drop                       48 users per Sector / 200 iterations

    UE Scheduling/Traffic Model     Round robin, full buffer

    Simulation Window               1 TTI

Ref.1: TR 36.814 v0.3.1 (2008-09), Further Advancements for E-UTRA, Physical Layer Aspects, 3GPP TR 25.942, 3GPP R1-084026
   Power Control in REC (3)
   Parameter Configuration in a Macro Cell Scenario

• Rationale is based on the
cell capacity and cell
coverage with considering
the corresponding average
interference over thermal
(IoT) level in the system
adopted from Ref.2

                                                                                    Cell coverage prioritized            Cell capacity prioritized
• Acceptable IoT level is
decided according to the                          Po & Alpha                                 -83 dBm & 1                       -42 dBm & 0.6
eNB receiver dynamic                              Average IoT                                   5.4 dB                              5.1 dB
range (see Appendix A)                            Cell capacity                               9354 kbps                          11032 kbps
                                                  Cell coverage                               3757 kbps                           3382 kbps

Ref.2: C. Castellanos, D. L. Villa, C. Rosa, I. Z. Kovacs, F. Frederiksen, and K. I. Pedersen, ‘’Performance of Fractional Power Control in UTRAN LTE
Uplink’’, The 2008 IEEE, ICC, Beijing, China, May 2008
                                                    Disclaimer: Resulting Po values are not same with the Ref.2 due to that shadowing is not considered.
Power Control in REC (4)
Suboptimal Settings for REC

   Macro cell scenario parameter configurations are named as full
    compensation power control (FCPC) and fractional power control
    (FPC) according to coverage and capacity priorities respectively
   The eNB-only deployment with optimal parameter settings for cell
    capacity prioritized scenario by FPC was assumed as reference case
    for the performance evaluation in REC scenario. Notations are as
       FPC: optimal parameter setting for fractional power control in eNB-only
       FCPC (eNB and RN): optimal parameter setting for FCPC in eNB-only
        deployment is applied in relay based deployment both at eNB and RN
       FPC (eNB and RN): optimal parameter setting for FPC in eNB-only
        deployment is applied in relay based deployment both at eNB and RN
Power Control in REC (5)
Results of Suboptimal Settings

                                                                          • Very high throughput at RNs
                                                                          • FCPC outperforms FPC up to
                                                                          50% ile
                                                                              • 80 % of the UEs connected
                                                                              to eNB experience higher
                                                                              throughput compared to FPC

                                                                          • FPC boosts the performance
                                                                          of UEs served by RNs

                                                                          • Do we need high capacity at
                                                                              • it should be noted that an
                                                                              ideal relay link is assumed (see
                                                                              Appendix B)

                                                                          • Parameter settings should be
  CDF of Throughput per UE at sector for FCPC (eNB and RN) vs. FPC
 (eNB and RN) in 1 tier (9 RNs deployed at the cell edges) REC scenario   re-adjusted to achieve an
                                                                          optimal performance
Power Control in REC (6)
Analysis of Po at eNB and RNs

• Analysis of Po at eNB in REC => Optimum cell edge performance can be maintained with
suboptimal settings found in eNB-only scenario

• Analysis of Po at RNs in REC => Feasible SINR threshold at RNs (-15 dBm), 12 dB lower
Po value can be used in FPC case
• Po value can be set as small as possible for the UEs served by RNs in order to improve
the performance of UEs served by eNBs
    Power Control in REC (7)
    Analysis of No Power Control at RNs

    Performing a power control scheme might still be regarded as an extra
     overhead at RNs
        No power control by considering fixed maximum allowed transmit power for UEs at RNs
        Scheme maintains the SINR performance of the cell edge users connected to RNs with a
         fixed maximum Tx power
            leads to higher throughput for the cell center UEs at RNs
            simpler RN design without penalizing the UEs served by eNB

                                                                            • 18 dBm illustrates
                                                                            similar performance to
                                                                            • 15 dBm results in better
                                                                            performance for the UEs
                                                                            at eNB
Power Control in REC (8)
Power control with Maximum Allowed Tx Power Setup

• Po configuration with fixed maximum allowed transmit power can be still re-
adjusted to reduce the experienced high throughput for the UEs served by RNs
and enhance the UEs served by eNBs

                                                              • 5% ile user throughput
                                                              improved by
                                                                   • 9 % for FCPC
                                                                   • 25 % for FPC
                                                              • Average user throughput
                                                              improved by
                                                                   • 17 % for FCPC
                                                                   • 40 % for FPC
                                                              compared to non-adjusted
                                                              suboptimal settings
    Power Control in REC (9)
    Results of Optimized Parameter Settings

    It is observed that FPC outperforms FCPC after parameter

    For the UEs at eNB in the REC
        FPC provides:
            70 % better cell edge user throughput (5 %ile) than eNB-only
            55 % better average user throughput than eNB-only

    For the same cell edge performance FPC provides:
        23 % better average user throughput at eNB than FCPC
        13 % better average user throughput at RNs than FCPC
        15 % better average user throughput at sector than FCPC
Summary & Conclusions
   Relaying is a promising solution for the demands of
   REC provides performance enhancement in the cell
    edge throughput and the system capacity compared to
    Macro cell scenario
   Standardized LTE UL Power control scheme is feasible
    to use in REC scenarios
   Parameter optimization and transmit power setup is
    important to achieve optimal performance
   FPC outperforms traditional FCPC with an appropriate
    parameter configuration and transmit power setup in
    REC scenarios
Appendix A
eNB Receiver Dynamic Range vs. Average IoT

    Assuming a maximum allowed receiver dynamic range of 35 dBm, compensation
     factors lower than 0.6 do not seem suitable to use because of non-acceptable eNB
     receiver dynamic ranges
    Appendix B
    Relay Link Overhead

   This study assumes an ideal relay link (can be maintained via Microwave transmission)
   However, a possible resource allocation scheme is also studied
to see the overhead that is introduced by relay link given as in Fig.1:
where half duplex transmission is applied to define the signal          f
                                                                          UE - eNB
                                                                                   RN - eNB
reception between direct link and relay link.                             UE - RN

  End-to-end user throughput is calculated according to
a minimum formula given as:                                                                 t

                                                          Relay Link Throughput per RN     Fig.1
e 2eUE Throughput  min(Access Link Throughput per UE ,                                )
                                                          Number of UEs served by RN

   It is observed that excessive user throughput
experienced from the access link is limited by the relay link
   Described resource allocation scheme only improves
the cell edge users while it does not increase the average
user throughput and system capacity compared to an
eNB-only scenario
         This can be achieved with bandwidth scalability of 100 MHz by LTE-A
          and more efficient resource allocation and frequency reuse scheme