doc.: IEEE 802.11-11/1170r0November 2009
M d li h h i d d 60 GHModeling the human induced 60 GHz channel dynamics
Date: 2009-11-17Authors:
Name Affiliations Address Phone email Martin Jacob Technische
UniversitätSchleinitzstr. 22 D 38092
+495313912451
Universität Braunschweig
D-38092 Braunschweig
51
Sebastian Priebe Technische Universität B h i
Schleinitzstr. 22 D-38092 B h i
Braunschweig BraunschweigThomas Kürner Technische
Universität Braunschweig
Schleinitzstr. 22 D-38092 Braunschweig
+495313912416
t.kuerner@ tu-bs.de
Submission Thomas Kürner, TU BraunschweigSlide 1
doc.: IEEE 802.11-11/1170r0November 2009
Abstract
Th t f l ti ith bl ki f i l th ( h lThe support of people motion with blocking of signal paths (channeldynamics) is currently an open issue in the TGad channel model [1].To overcome this we present simulations of the time-variant 60 GHzchannel based on ray tracing, a random walk model and a diffractionmodel. The results could be easily included in the channel modeldocument [2] as a further statistic.
Submission Thomas Kürner, TU BraunschweigSlide 2
doc.: IEEE 802.11-11/1170r0November 2009
Methodology2 5
3
Ray Tracing
0.5
1
1.5
2
2.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.50
Random Walk Model
0.06
0.08
0.1
DF
1st order reflections from walls
-0.2 0 0.2 0.4 0.6 0.8 1 1.20
0.02
0.04
Ed/E0
PD
Ch l St ti tiDiffraction Model
Set of time variant channel realizations
Channel Statistics
Submission Slide 3 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Diffraction Model• Assuming Person as „Multiple Edges“
)1(E
Assuming Person as „Multiple Edges• Electric field strength calculation based
on Fresnel Integral:
21
2
0
)(2
)exp(2
)1(
ddhmit
dttjjEEd
21
21
ddhmit
Ed
Verification of Model byMeasurements from [3]
• Loss:0
log20)(EEdBG d
d
• Good agreement between this simple model and measurements
Multiple Edge Model
Submission
Multiple Edge Model
Slide 4 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Ray Tracing Simulations
• STA-AP Scenario from [2]
AP iti fi d3
• AP position fixed
• STA positions uniformly distributed on 2
2.5
APthe table
Simulated:1071 STA positions0.5
1
1.5
STA
• Result: Each ray represents one cluster (cf. [2]) 1071 CIR for the static scenario
0 0.5 1 1.5 2 2.5 3 3.5 4 4.50
1071 CIR for the static scenario Categorized Rays (Table 2 from [2])
LOS 1st order
Submission
2nd order
Slide 5 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Random Walk Model
• Simulation Parameters 1 Person walking around the table
d lk i h f di i Random Walk with preference direction Speed v =1m/s Step size: 60 cm Person assumed as cuboid:
0.45m x 0.15m x 1.70m
Time resolution: 10 ms Simulation Time: 30 s
A l diff ti d l t th t i lt Apply diffraction model to the ray tracing results
3001 different time-variant CIRs for each of the 1071 STA
Submission
positions
Slide 6 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Simulation Example for Random Walk
Submission Slide 7 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Example Results
Temporal variation of the diffraction Gain of the four 1st order reflected raysfor a fixed STA position on the table
Submission Slide 8 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Statistics1. Number of blocked clusters
• Based on pure „blockage approach“• Full blockage if ray intersects personFull blockage if ray intersects person• 2 categories (1st & 2nd order)
2. Blockage probabilities for single clusters Based on diffraction model Full blockage if ray is attenuated by more than 2 dBFull blockage if ray is attenuated by more than 2 dB 2 categories (1st & 2nd order)
3. Amplitude distribution Based on diffraction model
Attenuation due to diffraction
Submission
Attenuation due to diffraction
Slide 9 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Number of blocked clusters - 1
• 1071 x 3001 = 3,214,071 channel realizations• In 99.99 % of the realizations the LOS link was NOT blocked!
• 1st order reflections from walls
1 t d fl ti f ll
0.6
0.8
1
F
1st order reflections from wallsNumber of blocked
ClustersProbability
0 80.4 %
0
0.2
0.4PD
F
1 13.2 %
2 6.4 %
3 0 %0 1 2 3 4
0
Number of blocked clusters 4 0 %
Submission Slide 10 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Number of blocked clusters - 2
• 2nd order reflections from two walls
Number of blockedClusters
Probability
0 71.6 %2nd order reflections from two walls
1 10.5 %
2 6.2 %
3 3.8 %0 4
0.6
0.8
DF
4 0.3 %
5 0 %
6 6.4 %0 1 2 3 4 5 6 7 8
0
0.2
0.4
PD
7 0.7 %
8 0.5%
0 1 2 3 4 5 6 7 8Number of blocked clusters
Submission Slide 11 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Blockage Rate of single clusters
• The blockage rate of single clusters is calculated as the ratio of the duration ofblockage and the simulation time (relative blockage time)
• In 99.99 % of the realizations the LOS link was NOT blocked!
0.51st order reflections from walls
0.42nd order reflections from walls
• For the four 1st order reflections and the eight 2nd order reflections pdfs are given:
0 2
0.3
0.4
PD
F
0.2
0.3
PD
F
0 0.05 0.1 0.15 0.2 0.250
0.1
0.2
0 0.1 0.2 0.3 0.4 0.50
0.1
Submission
relative blockage time relative blockage time
Slide 12 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0
A li d di ib i f d hNovember 2009
Amplitude distribution of attenuated paths0.2
1st order reflections from walls0.1
1st order reflections from walls
0 05
0.1
0.15
PD
F
0.04
0.06
0.08
PD
F
-40 -30 -20 -10 0 100
0.05
Ed/E0
-0.2 0 0.2 0.4 0.6 0.8 1 1.20
0.02
Ed/E0
2 d d fl ti f ll2 d d fl ti f ll
0 1
0.15
0.2
DF
2nd order reflections from walls
0.06
0.08
0.1
DF
2nd order reflections from walls
-40 -30 -20 -10 0 100
0.05
0.1
PD
-0.2 0 0.2 0.4 0.6 0.8 1 1.20
0.02
0.04PD
Ed/E0Ed/E0
• Only attenuated paths taken into account (Ed/E0 <-2 dB)• Otherwise the amplitude distribution would be nearly Dirac shaped at Ed/E0 =1 (0 dB),
Submission
p y p d 0 ( ),because the single path is not blocked most of the time
Slide 13 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
Conclusion
We propose to introduce the channel dynamics into theTGad channel document based on the presentedpsimulation results. From this data it is possible toextract statistics about the total number of blockedclusters, about the blockage rate of single clusters orabout the amplitude distribution of single clusters.These statistic can be included in the channel model inThese statistic can be included in the channel model inan adequate way.
Submission Slide 14 Thomas Kürner, TU Braunschweig
doc.: IEEE 802.11-11/1170r0November 2009
References
• [1] E. Perahia: TGad Task Group Document Open Items IEEE802.11-09/1108r0, October 2009
• [2] A. Maltsev et al.: Channel Models for 60 GHZ WLAN Systems, IEEE802.11-09/0334r3, July 2009
• [3] M. Jacob et al.: Human Body Blockage - Guidelines for TGad MAC[3] M. Jacob et al.: Human Body Blockage Guidelines for TGad MAC Development, IEEE802.11-09/1169r0, November 2009
Submission Slide 15 Thomas Kürner, TU Braunschweig