Project: IEEE P Working Group for Wireless Personal Area Networks N

Transcription

1 Project: IEEE P82.15 Working Group for Wireless Personal Area Networks N (WPANs( WPANs) Title: [Merging two-path and S-V models for LOS desktop channel environments] Date Submitted: [July, 26] Source: [Hirokazu Sawada, Yozo Shoji, Chang-Soon Choi, Katsuyoshi Sato, Ryuhei Funada, Hiroshi Harada, Shuzo Kato, Masahiro Umehira, and Hiroyo Ogawa] Company [National Institute of Information and Communications Technology] Address [3-4, Hikarino-Oka, Yokosuka, Kanagawa, , Japan] Voice:[ ], FAX: [ ], [sawahiro@nict.go.jp] Re: [] Abstract: [This contribution describes update of the generic channel model merging two-path and S-V models.] Purpose: [Contribution to mmw TG3c meeting.] Notice: This document has been prepared to assist the IEEE P It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P Slide 1

2 Agenda Channel model for LOS desktop environments Proposal of TSV model Measurement procedure and results Extracted TSV model parameters Slide 2

3 Importance of channel model for LOS desktop U16 U16 U16 U3 U3 U3 LOS desktop is one of useful channel environments for TG3c Important to develop channel model for LOS desktop Slide 3

4 What is suitable channel model for LOS desktop? Two-path model is suitable to express LOS desktop environment (6/19) Developing statistical two-path model and merging with S-V model was proposed for TG3c generic channel model (6/228) This model is named TSV model (Triple SV: Shoji, Sawada, Saleh and Valenzuela model) Slide 4

5 Results: Path loss in two-path response -7 Path loss [db] Distance: 3m Tx height: 15mm Meas. Two-path Free-space Tx,Rx height: 15mm Receiver antenna height [mm] Path loss vs Rx height Path loss vs distance Two-path model is suitable to express propagation phenomena in LOS desktop environment Slide 5

6 Proposed TSV model TSV model = Statistical two-path model + S-V model h L 1 M l 1 () t = β δ () t + αl, m δ ( t Tl τ l, m ) δ ( ϕ Ψl ψ l, m ) Gr (, Ψl + ψ l, m ) l= m= Statistical two-path response (LOS desktop.) Fixed impulse response (LOS office, residential etc) Relative power β Ω Rician factor (ΔK) S-V model response Γ,Λ,γ,λ Time of Arrival Slide 6 Refer to Appendix A for each parameter

7 Purpose of measurement To confirm the validity of TSV model in LOS desktop environments To extract TSV model parameters Slide 7

8 Measurement conditions Instrument Center frequency Bandwidth Time resolution Distance resolution # of frequency points Frequency step Times of average HP851C VNA 62.5 GHz 3 GHz.333 ns 19.1 cm MHz 128 times Time resolution and distance resolution were determined by measurement bandwidth Slide 8

9 Measurement conditions (cont ) Antenna: Conical horn antenna Polarization: Vertical Beam-width: Tx:3 and Rx 3, Tx:6 and Rx6 Conical horn antenna Beam-width 3 deg Conical horn antenna Beam-width 6 deg Slide 9

10 Measurement environment Transmitter Receiver Rotations Small conference room: 6.4 m 7.4 m Ceiling height: 2.7 m Surrounding: metallic wall, glass window Floor: Plaster board covered with carpet Furniture: Wooden desk, chair, computer, LCD TV, white board Receiver was rotated from to 36 degree in 5 degree step Slide 1

11 Measurement environment Absorber Antenna Rotations Tx side Rx side Receiver was not put on the desk due to large rotator size Calibration was done at 1 m distance Slide 11

12 AoA measurement environment (Two-path) Metal based inner wall 64 unit: cm 6 42 Wooden desk Reflection from the desk Tx 6 No desk area 6 Chair Rx 5m 32 Steel cabinet 15 Glass 1m 2m 3m 4m Room height: Pillar 77 LCD TV White board Antenna height from desktop Tx = 15 cm Rx = 15 cm Slide 12

13 AoA measurement environment (Non-two-path) Metal based inner wall 64 unit: cm Tx 6 42 Wooden desk No reflection from the desk m No desk area 6 Chair Rx Steel cabinet 15 Glass 2m 3m 4m Room height: Pillar 77 5m LCD TV White board Antenna height from desktop Tx = 15 cm Rx = 15 cm Slide 13

14 AoA measurement environment (Spatial) 32 Metal based inner wall Tx Wooden desk No desk area 1 cm grid (5 5=25 points) 6 unit: cm Chair 32 To characterize small-scale fading Only 3m transmission 25 Steel cabinet 15 Glass Room height: m Rx Pillar 77 LCD TV White board Antenna height from desktop Tx = 15 cm Rx = 15 cm Slide 14

15 AoA measurement environment (Two heights) Metal based inner wall 6 42 unit: cm 64 Wooden desk Only 3m transmission Tx Steel cabinet 15 Glass 6 No desk area 96 3m Room height: 27 Rx Pillar 77 6 Chair LCD TV White board Antenna height from desktop Tx = 17, 14.3 cm high- and low- Rician cases Rx = 15 cm Slide 15

16 Measurement results Slide 16

17 Relative power [db] July 26 Relative power [db] PDPs for two different antenna heights Two-path response ΔK = 24.1 db Antenna height Tx: 17 mm Rx: 15 mm Beam width: 6 deg Distance: 3m S-V cluster Two-path response ΔK = 5.1 db Antenna height Tx: 143 mm Rx: 15 mm Beam width: 6 deg Distance: 3m S-V cluster Time of arrival [ns] (a) High Rician factor Side-lobe effect of window function in IFFT Slide Time of arrival [ns] (b) Low Rician factor Δ K has 19dB dynamic range TSV model is well expressing LOS desktop

18 TSV model parameters to be extracted σ1 ln(h 2 (t)) AoA Ω σ2 γ Γ Γ : cluster decay factor 1/ Λ : cluster γ : ray decay factor 1/ λ : ray arrival rate arrival rate σ : cluster lognormal standard deviation 1 σ : ray lognormal standard deviation 2 σ φ : Angle spread of ray within cluster Two-path Response σ φ 1/Λ 1/λ ToA Ω (Laplace distribution) : Average power of of the first cluster the first ray S-V parameters and Ω are required for TSV model Slide 18

19 Extracted TSV model parameters Parameter Tx:3 Rx:3 Tx:6 Rx:6 TSV Model Ω (D) [db] 4.44 D D-98 Parameter Tx:3,Rx:3 Tx:6,Rx:6 Γ [ns] S-V model oriented parameter 1/Λ [ns] γ [ns] Slide 19 1/λ [ns] σ 1 cluster σ 2 ray Example of ΔK in measurement Max.ΔK[dB] Two-path@3m Variable Min.ΔK[dB] σ φ [deg] Non-two path@3m ΔK[dB] Constant Number of cluster 3 3

20 Calculation of ΔK Extracted TSV parameter and calculated two-path response were used d 1 h Uniform 1 d 2 h 2 Uniform D Uniform Assumption: Transceiver position parameters have uniform distribution Slide 2

21 CDF Calculated CDF of ΔK Tx:6 Rx:6 Distance [m] h h 1 2 Uniform(,.3) Uniform(,.3) D Uniform ( μ D.3, μ Unit [m] D +.3) ΔK [db] High ΔK can be usually obtained in short range communication Slide 21

22 TSV model for LOS residential environments TSV model can be applied in any LOS environment TSV model parameters for LOS residential environments can be obtained since we have measurement data (6/12) Tx Rx Residential environment Slide 22

23 Large ΔK for LOS environment with high directivity antenna Relative amp [db] LOS Residential (NICT data) Tx: 15 deg Rx: 15 deg Tx ΔK=5dB Rx ToA [ns] Antenna beamwidth Tx:15, Rx:15 Tx:3, Rx:15 Tx:6, Rx:15 Tx:Omni, Rx:15 Slide 23 Max. ΔK[dB] ΔK is very large in LOS environments with directional antenna The effect of S-V model response is decreasing in LOS environment

24 Conclusion TSV model is well expressing LOS desktop environments Characteristics of variable ΔK were clarified TSV model with extracted parameters is now available TSV model can be applied for any LOS environment MATLAB code for TSV model will be available soon Slide 24

25 CIR: 2 μ D 2π 2h1h 2 β = Gt1Gr1 + Gt 2Gr 2Γ exp j D λ f D Path number of G ti and G (1: direct, 2 : refrect) Appendix A: Definition of TSV model h (Complex impulse response) α L 1 M l 1 () t = β δ () t + αl, m δ ( t Tl τ l, m ) δ ( ϕ Ψl ψ l, m ) Gr (, Ψl + ψ l, m ) 2 = Ω l, m Two-path response ri l e T Γ e l= m= τ l, m γ, α l, m p p [ π ) Uniform,2 Arrival rate: Poisson process ( T T ) = Λ exp[ Λ( T T )], l l 1 l 1 [ ( )], m > ( τ l τ l, ( m 1) ) = λ exp λ τ l τ l, ( m 1) l l > t: time[ns] δ( ): Delta function l = cluster number, m= ray number in l-th cluster, L = total number of clusters; M l = total number of rays in the l-th cluster; T l = arrival time of the first ray of the l-th cluster; τ l,m = delay of the m-th ray within the l-th cluster relative to the firs path arrival time, T l ; Ω = Average power of the first ray of the first cluster Ψ l Uniform[,2π); arrival angle of the first ray within the l-th cluster ψ l,m = arrival angle of the m-th ray within the l-th cluster relative to the first path arrival angle, Ψ l Two-path parameters (4) S-V parameters (7) D h h 1 2 Γ Uniform : Distance between Tx and Rx Uniform : Height of Uniform : Height of Tx Rx 1: Reflection coefficient (incident angle π 2) Γ : cluster decay factor 1/ Λ : cluster γ : ray decay factor 1/ λ : ray σ : cluster lognormal standard deviation 1 σ : ray lognormal standard deviation 2 σ φ arrival rate arrival rate : Angle spread of ray within cluster (Laplace distribution) K Antenna parameters (2) Gt Gr = L 1 M 1 ( θ, φ) : Antenna gain of Tx ( θ, ι) : Antenna gain of Rx l l= m= Rician factor (1) α 2 l, m δ ( t T τ ) δ ( ϕ Ψ ψ ) G (, Ψ + ψ ) l l, m 2 β l l, m r l l, m Slide 25

26 ln((α l, /α, ) 2 ) y = x Γ = T l [ns] Appendix B: Results of data analysis ln(1-cdf) y = x 1/Λ = 21.1 [ns] T l [ns] Normalized relative power of ray y = x γ = 17.2 [ns] Relative delay of ray [ns] 1-cdf Antenna beamwidth Tx: 6 deg, Rx: 6 deg y=-.373x 1/λ = 2.68 [ns] τ l,m [ns] Cluster decay factor (Γ) Cluster arrival rate (1/Λ) Ray decay factor (γ) Ray arrival rate (1/λ) cdf pdf ( x) y = e σ φ = 38.1 [deg] AoA of cluster [deg] AoA [deg] Angle of arrival in cluster ( Uniform) Angle spread of ray (σ φ ) Standard deviation of cluster (σ 1 ) Standard deviation of ray (σ 2 ) Slide 26

27 Appendix C: Averaged power of the first ray of S-V response Ω [db] Ω = 3.46 D wall Path #1 Path #2 Tx Rx Rx Distance [m] Directivity d 1 d 2 Ω increases due to distance, because directional antenna is used in transmitter Conventional S-V model does not consider this effect Slide 27