Channel Modeling for 60 GHz WLAN Systems by L7VpJy

VIEWS: 23 PAGES: 20

									July 2008                                           doc.: IEEE 802.11-08/0811r1


     Channel Modeling for 60 GHz WLAN Systems
                               Date: 2008-07-14
Authors:

Name             Affiliations Address           Phone           email
Alexander        Intel       Turgeneva str., 30, +78314162461   alexander.maltsev@intel.com
Maltsev                      Nizhny Novgorod,
                             603024, Russia
Ali Sadri        Intel                                          ali.s.sadri@intel.com

Roman            Intel                                          roman.maslennikov@intel.com
Maslennikov
Alexei Davydov   Intel                                          alexei.davydov@intel.com

Alexey Khoryaev Intel                                           alexey.khoryaev@intel.com




Submission                            Slide 1                         Alexander Maltsev, Intel
 July 2008                             doc.: IEEE 802.11-08/0811r1

                          Abstract

• This presentation discusses general characteristics of the 60
  GHz radio propagation channel, provides an overview of the
  measurements results and the available channel models, and
  proposes approaches to the channel modeling for IEEE VHT 60
  GHz




 Submission                  Slide 2                Alexander Maltsev, Intel
 July 2008                              doc.: IEEE 802.11-08/0811r1

                           Agenda

• Characteristics of the 60 GHz propagation channel
• IEEE 802.15.3c WPAN channel models review
• Possible approaches to IEEE VHT channel modeling
• Conclusion




 Submission                   Slide 3                Alexander Maltsev, Intel
 July 2008                                       doc.: IEEE 802.11-08/0811r1
     General Characteristics of 60 GHz Propagation
                       Channel
• The propagation characteristics at the 60 GHz band are significantly
  different from that for the current WPAN / WLAN bands of 2 - 5 GHz.
• The main difference is that the 60 GHz propagation has a quasi-optical
  nature.
• The 60 GHz propagation loss (under the same TX and RX antenna gains) is
  20 to 30 dB higher than for 2 – 5 GHz band.
• The diffraction effects (propagation of the EM field behind the obstacle) are
  significantly smaller in comparison with 2 – 5 GHz band (shadowing zones
  are very sharp in 60 GHz).
• The penetration loss for 60 GHz band is also higher than for
  2 – 5 GHz band.
• The cross polarization level of the received signal for the 60 GHz band is on
  average significantly below than for 2 – 5 GHz and the impact of the
  polarization may be more important for 60 GHz than for 2 – 5 GHz.
 Submission                          Slide 4                  Alexander Maltsev, Intel
 July 2008                                       doc.: IEEE 802.11-08/0811r1

     General Characteristics of 60-GHz Propagation
                  Channel (Cont’d)
• The EM field for 2 - 5 GHz signal is composed of multiple waves coming
  from primary signal source and multiple secondary sources – diffracted and
  reflected (multiple times) waves. The EM field has a complex structure with
  no preferable directions of arrival and departure of the communication
  signals. The multiple antenna algorithms are focused on exploitation of the
  spatial diversity of the EM field in different points.
• Due to specific properties of 60 GHz EM field listed above it has a structure
  consisting of a few rays coming from the direct path (if available) and from
  several main reflectors with the directions of arrival and departure very
  close to that predicted by the ray tracing (geometrical optics) laws. So the
  antenna processing for 60 GHz should be focused on the spatial filtering of
  one or few rays to maximize the received signal power. High directional
  antennas may do that.


 Submission                          Slide 5                   Alexander Maltsev, Intel
July 2008                                                   doc.: IEEE 802.11-08/0811r1

       Typical EM Field Structure for 2 – 5 GHz [1]




 •    The EM field has complex quasi random structure consisting of multiple running and standing
      waves overcoming obstacles due to strong diffraction and reflection effects.
 •    The field structure is ideal for statistical description and statistical channel modeling [2]

Submission                                   Slide 6                        Alexander Maltsev, Intel
    July 2008                                           doc.: IEEE 802.11-08/0811r1

60-GHz Propagation Measurement Results (Office) [3]




•   The 60 GHz received signal is mainly a combination of the direct path and first-order
    reflected signals, which may be predicted by the image-based ray-tracing procedure
    Submission                            Slide 7                     Alexander Maltsev, Intel
July 2008                                 doc.: IEEE 802.11-08/0811r1

  60-GHz Propagation Measurement Results (Aircraft) [4]



                                         Receiver




                         •    TX antenna – open waveguide (5 dBi), RX
                              antenna – omni-directional in azimuth (2
                              dBi)
                         •    The signal consists of the LOS component
                              and multiple reflected signals bounced from
                              the walls of the aircraft (the estimated
                              reflection coefficient is about -8 … -10 dB)
Submission                   Slide 8                     Alexander Maltsev, Intel
  July 2008                             doc.: IEEE 802.11-08/0811r1

60-GHz Propagation Measurement Results (Home and Office) [5]
                          Residential                 Office




                                                                                Measured
• The good matching
  between measured
  cluster    positions
  and       simulated
  cluster    positions
  using ray tracing




                                                                                Simulated
  was obtained




  Submission                  Slide 9                Alexander Maltsev, Intel
 July 2008                                        doc.: IEEE 802.11-08/0811r1

              IEEE 802.15.3c Channel Modeling
• The channel modeling was done by the IEEE 802.15.3c group in Q1’2006 –
  Q1’2007.
• Different application scenarios for WPAN systems were identified and the
  goal was set to develop the channel models for all considered scenarios.
• The statistical channel models developed by the IEEE 802.15.3c group
  included the following features:
   – LOS and NLOS components;
   – Based on the generalized Saleh-Valenzuela channel model with
      clustering in both the time and angular domain;
   – Clusters arrival time and intra-cluster rays arrival time are two Poisson
      processes;
   – The distributions of the clusters and intra-cluster rays amplitudes are log-
      normal;
   – The distribution of the clusters angle-of-arrival (AoA) is uniform. AoAs
      of different clusters are independent. The distribution of rays AoA inside
      the cluster is Gaussian;

 Submission                          Slide 10                   Alexander Maltsev, Intel
 July 2008                                        doc.: IEEE 802.11-08/0811r1

              IEEE 802.15.3c Library Channel Model
• The Intel team was working in the IEEE 802.15.3c group on the
  development of the library channel model [6 – 12]
• The library channel model was developed based on the high quality
  experimental data obtained by the German company IMST.
• The measurements were performed in the library room with tables, chairs
  and metal bookshelves with books
• 3 types of RX antennas (horn, wideband dipole array antenna, biconical)
• Fixed TX lens antenna position at the suspended ceiling, RX measurements
  range ~2-5m
• Time resolution is 1/960MHz ≈ 1ns
• Two types of virtual uniform antenna arrays for direction of arrival analysis
   – 501x1 uniform linear array with 1mm antenna spacing (LOS scenarios)
   – 301x51 uniform planar antenna array with 1mm antenna spacing (edge
      scenario)
 Submission                          Slide 11                  Alexander Maltsev, Intel
 July 2008                                           doc.: IEEE 802.11-08/0811r1
Library Channel Model – Measurements Scenarios Plan
                                     LOS                       NLOS




Edge          Transmitting antenna              Receiving antennas




 Submission                          Slide 12                        Alexander Maltsev, Intel
     July 2008                                             doc.: IEEE 802.11-08/0811r1
         Example of 60 GHz Channel for Library Measurements
             (Time-Angular Energy Distribution Function)




•   The channel has a “specular” structure with direct ray and several strongly localized
    reflected rays (clusters) in time/angular dimensions for delays < 50 ns and angles < 700
     Submission                             Slide 13                     Alexander Maltsev, Intel
July 2008                                                                               doc.: IEEE 802.11-08/0811r1

      Generalized Saleh-Valenzuela Channel Model
                                         L 1 Kl 1
         ht ,   K LOS   t ,     k ,l t  Tl   k ,l    l   k ,l 
                        
                     l 0 k 0 
                           LOS                                                                   MPC

  •    L – Number of clusters
  •    Kl – Number of MPC in the lth cluster                                            40



  •    KLOS – LOS K factor
                                                                                        35


                                                                                        30
  •    KMP – Cluster K factor                                                                                   KLOS




                                                                Relative Power, [dB]
                                                                                        25
  •    αk,l – MPC complex amplitude
                                                                                        20
  •    Tl – Time of arrival of lth cluster
                                                                                        15
  •    τk,l – Relative time of arrival for kth MPC                                                                                                KMP
                                                                                        10
       within lth cluster
                                                                                         5
  •    θk,l – Relative direction of arrival for kth MPC
                                                                                         0
       within lth cluster                                                                    0    10       20    30      40
                                                                                                                      Delay, [ns]
                                                                                                                                    50       60    70       80


  •    Θl – Direction of arrival of lth cluster                                                                                                         2j k ,l
  •    δ(·) – Delta function                                                            k ,l                   l
                                                                                                                 
                                                                                                                                         k ,l    e
                                                                                                                                        
                                                            Com plex amplitude                             Cluster amplitude MPC amplitude


Submission                                       Slide 14                                                             Alexander Maltsev, Intel
        July 2008                                                                     doc.: IEEE 802.11-08/0811r1

                         Extracted Channel Model Parameters
Parameter      Γ, [ns]   γ, [ns]   Λ, [ns-1]   λ, [ns-1]   Klos, [dB]   KMP, [dB]   σAS, [deg]   σ1, [dB]   σ2, [dB]     τ, [ns]   Δ, [dB]
Value            12         7        0.25          4           8           -13          10           5          6          30          11
                                                                                                  Log-        Log-
Distribution    N/A       N/A      Poisson     Poisson        N/A          N/A      Gaussian     Normal      Normal       N/A          N/A

         •   Inter-cluster Power Delay Profile parameters
               – Exponential PDP:         K factor KLOS = 8 [dB], cluster decay Γ = 12 [ns]
               – Flat-Exponential PDP: K factor KLOS = 8 [dB], cluster decay Γ = 12 [ns],
                                                     Δ = 11 [dB], τ = 30 [ns]
         • Inter-cluster DoA – Uniform distribution
         • Intra-cluster DoA – Gaussian (Angles Spread (AS) σAS = 10 [0])
         • Inter-cluster ToA / Intra-cluster ToA – Poisson (Λ = 0.25 [ns-1] / λ = 4 [ns-1])
         • Cluster / Ray amplitude – Log-normal (σ1 = 5 [dB] / σ2 = 6 [dB])
         • Intra-cluster K factor KMP = -13 [dB]
         • Intra-cluster ray decay γ = 7 [ns]
         Note: Two approximations of cluster PDP are proposed:
         Exponential PDP - for TX omni-directional antenna mounted at the suspended ceiling.
         Flat-exponential PDP - for TX beam-shaped antenna mounted at the suspended ceiling.

        Submission                                                  Slide 15                                Alexander Maltsev, Intel
 July 2008                                     doc.: IEEE 802.11-08/0811r1

              IEEE 802.15.3c Channel Models Summary

• Some work done in the IEEE 802.15.3c channel modeling may be reused for
  IEEE VHT WLAN channel models.
• The reuse is complicated by the fact that no raw channel measurement
  results (except for the IMST data used in the library model) are available
  now.
• The mandatory extensions of the IEEE 802.15.3c channel models should be
  the support of the both the TX and RX steerable directional antennas and
  polarization characteristics [12].
• The library channel model is the most suitable and complete (from IEEE
  802.15.3c channel models) for the WLAN environment – the measurements
  are done in a large room with access point sitting near the ceiling and
  transmitting in quasi-omni mode and the receiver using “virtual” antenna
  array. The model and the measurements data may be reused for the IEEE
  VHT channel modeling.
 Submission                        Slide 16                 Alexander Maltsev, Intel
July 2008                                       doc.: IEEE 802.11-08/0811r1

             Requirements for VHT Channel Model
 • The following basic requirements are proposed for IEEE
   VHT channel model development:
       – Provide accurate space-time characteristics of the 60 GHz
         propagation channel for main usage models.
       – Support the beamforming with the steerable directional antennas
         at both TX and RX sides.
       – No limitation on the antenna type (i.e. antenna arrays, sector-
         switching antennas, non-steerable antennas) and antenna
         parameters.
       – Include the polarization characteristics of the antennas and
         signals. (Additional experimental investigations should be done).




Submission                          Slide 17                 Alexander Maltsev, Intel
 July 2008                                         doc.: IEEE 802.11-08/0811r1

       Channel Modeling Approaches for IEEE VHT

• The following approaches may be considered for IEEE channel
  modeling:
     1. Statistical channel model – randomly generation of the channel
        realizations with the required statistical characteristics based on several
        input parameters (examples – IEEE 802.11n channel model [2], IEEE
        802.15.3c library channel model [10]).
     2. Use of experimental golden sets – to collect a number of experimental
        golden sets for the typical scenarios and use them for system simulation
        and performance evaluation.
     3. Combination of ray-tracing and statistical channel model – use the
        ray-tracing for the cluster parameters identification and generate intra-
        cluster rays distribution statistically.
• More than one approach may be used by the VHT group
 Submission                           Slide 18                   Alexander Maltsev, Intel
 July 2008                                    doc.: IEEE 802.11-08/0811r1

                            Conclusions

• The propagation characteristics of signals for 60 GHz frequency band are
  different from the 2 – 5 GHz band. The mmWave signal propagation has a
  quasi-optical nature. The main propagation mechanisms for 60 GHz are
  direct path (LOS) and reflections (NLOS).
• The IEEE 802.15.3c group has developed the channel models for 60 GHz
  WPAN scenarios. Some of those models may be reused for the VHT with an
  appropriate modification.
• The additional measurement campaigns have to be carried out for the WLAN
  scenarios to obtain the required statistical parameters of the 60 GHz
  channels.




 Submission                       Slide 19                 Alexander Maltsev, Intel
July 2008                                                     doc.: IEEE 802.11-08/0811r1

                                       References
1.  IEEE Standard 802.11a
2. V. Erceg et al. “TGn Channel Models”, IEEE 802.11 document 11-03/0940r4.
3. H. Xu, V. Kukshya, and T. S. Rappaport, “Spatial and Temporal Characteristics of 60 GHz Indoor
    Channels,” IEEE J. Sel. Areas Commun., vol. 20, no. 3, pp. 620–630, Apr. 2002.
4. M. Peter, W. Keusgen, A. Kortke, M. Schirrmacher, “Measurement and Analysis of the 60 GHz In-
    Vehicular Broadband Radio Channel”, Proc. of IEEE Vehicular Technology Conference, 2007 (VTC-2007),
    pp. 834 - 838
5. B. Neekzad, K. Sayrafian-Pour, J. Perez, J. S. Baras, “Comparison of Ray Tracing Simulations and
    Millimeter Wave Channel Sounding Measurements”, Proc. of the 18th Annual IEEE International
    Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'07)
6. IEEE 802.15-06-0022-01-003c, A. Sadri, A. Maltsev, A. Davydov “IMST data analysis ,Preliminary
    results”, January 2006.
7. IEEE 802. 15-06-0141-01-003c, A. Sadri, A. Maltsev, A. Davydov “IMST time-angular characteristics
    analysis”, March 2006.
8. IEEE P802.15-06-0201-00-003c, A. Davydov, A. Maltsev, A. Sadri “IMST Data Processing Methodology”,
    April 5, 2006.
9. IEEE 802.15-06-0230-00-003c, A. Davydov, A. Maltsev, A. Sadri “Resolving the Ambiguity in IMST
    Measurements”, May 2006.
10. IEEE 802.15-06-0302-02-003c, A. Davydov, A. Maltsev, A. Sadri “Saleh-Valenzuela Channel Model
    Parameters for Library Environment”, July 2006.
11. IEEE 15-07-0715-00-003c, A. Maltsev, R. Maslennikov, A. Khoryaev “Comments on CM3.1 golden set”,
    May 2007.
12. IEEE 15-07-0774-00-003c , A. Maltsev, A. Davydov “Generalization of TG3c channel models”, June 14,
    2007.
Submission                                    Slide 20                          Alexander Maltsev, Intel

								
To top