September 2002
doc.: IEEE 802.15-02/383 SG3a
Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs)
Submission Title: Comparison of S-V Channel Model to Empirical Data
Date Submitted: 08 September, 2002
Source: Marcus Pendergrass, Time Domain Corporation
7057 Old Madison Pike, Huntsville, AL 35806
Voice:256-428-6344 FAX: [256-922-0387], E-Mail: [email protected]
Re: Ultra-wideband Channel Models IEEE P802.15-02/208r0-SG3a, 17 April, 2002,
Abstract: Channel realizations generated by the S-V model are compared to measured channels. We show
that the S-V model is able to accurately reproduce population means for 3 channel parameters: RMS delay
spread, mean excess delay, and number of significant paths. In addition, a hueristic channel classification
scheme is proposed, and we show that the S-V model is able to produce exemplars of each channel class
except for one. We recommend that SG3a adopt a channel model consisting of a defined set of channel
realizations generated by the S-V model, augmented with additional channel realizations necessary to
capture the wide variety of channels that occur in typical WPAN application scenarios.
Purpose: The information provided in this document is for consideration in the selection of a UWB
channel model to be used for evaluating the performance of a high rate UWB PHY for WPANs.
Notice: This document has been prepared to assist the IEEE P802.15. 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 P802.15.
Submission
Slide 1
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Summary
• S-V Model can match measured channels in terms of the population
means of Mean Excess Delay, RMS Delay Spread, and Number of
Significant Paths for three scenarios:
• NLOS, 0 to 4 meters, office
• LOS, 0 to 4 meters, office
• NLOS, 4 to 6 meters, office
• Variation of measured channels is very wide, even within a single
scenario.
• A heuristic channel taxonomy is proposed.
• Single set of S-V parameters for each scenario is probably not sufficient
to capture the wide variation among channels from a given category.
• Recommendations
• channel model should consist of a defined set of channel impulse
response realizations
• make sure all channel types within a given scenario are represented
• multiple S-V parameter sets per scenario
Submission
Slide 2
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
S-V Model Can Match Population Means of Data
Measurements
S-V Model
Case:
Case:
1
Channel Parameter
2
3
Population means
1
2
3
Population means*
Mean excess delay
10.38
5.05
14.18
10.04
5.28
15.16
RMS delay spread
8.03
5.28
14.28
8.33
5.50
15.07
Number of Paths (85 % energy capture)
36.1
24.0
61.54
35.88
24.39
61.77
*data provided by Jeff Foerster
Case 1: NLOS, 0-4 meters, office
Case 2: LOS, 0-4 meters, office
Case 3: NLOS, 4-10 meters, office
Submission
Slide 3
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Heuristic Channel Taxonomy
Channel Type
Measurements
S-V model
Case:
Case:
1
2
3
1

Freespace
Freespace + diffuse multipath
Closely spaced dominant paths
Closely spaced dominant paths + strong late arriving cluster
LOS path not dominant, little diffuse multipath
LOS path not dominant + diffuse multipath
Strong late arriving cluster, little diffuse multipath
Strong late arriving cluster + diffuse multipath
Diffuse multipath, no single dominant path
2
3
?

 
   ?



 
   
   ?

 ?

 

? 
 
? 
?
Case 1: NLOS, 0-4 meters, office
Case 2: LOS, 0-4 meters, office
Case 3: NLOS, 4-10 meters, office
Submission
Slide 4
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
A Channel Menagerie
• The next several slides give visual depictions of selected channels from
each Channel Type described on the previous slide.
• Where possible, we show both measured and simulated channels for
each Channel Type.
• Purpose:
• illustrate the variety of channels in the measurement data
• indicate the ability of the S-V model to provide a reasonable
facsimile to these channels
• Measured channels have RED titles
• Simulated channels have GREEN titles
Submission
Slide 5
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Freespace
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 6
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Freespace + diffuse multipath
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 7
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Freespace + diffuse multipath
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 8
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Closely spaced dominant paths
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 9
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Closely spaced dominant paths
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 10
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Closely spaced dominant paths, strong late-arriving cluster
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 11
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Closely spaced dominant paths, strong late-arriving cluster
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 12
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
LOS path not dominant, little diffuse multipath
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 13
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
LOS path not dominant, little diffuse multipath
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 14
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
LOS path not dominant + diffuse multipath
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 15
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
LOS path not dominant + diffuse multipath
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 16
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Strong late-arriving cluster, little diffuse multipath
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 17
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Strong late-arriving cluster, little diffuse multipath
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 18
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Strong late-arriving cluster + diffuse multipath
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 19
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Strong late-arriving cluster + diffuse multipath
S-V CIR
S-V reconstructed waveform
95% energy
capture
85% energy
capture
Submission
Slide 20
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Diffuse multipath, no single dominant path
Measured data
spectrum
DSO Scan
CIR
Reconstructed
Scan
Submission
Slide 21
Marcus Pendergrass
Time Domain Corporation (TDC)
September 2002
doc.: IEEE 802.15-02/383 SG3a
Recommendations
• Channel model should consist of a defined set of at least 50 channel
impulse response realizations for each of the following 6 scenarios.
• NLOS, 0 to 4 meters, office
• LOS, 0 to 4 meters, office
• NLOS, 4 to 10 meters, office
• NLOS, 0 to 4 meters, residential
• LOS, 0 to 4 meters, residential
• NLOS, 4 to 10 meters, residential
• Channel realizations should include all types of channels found in the
measured data for each scenario (see channel taxonomy, slide 3).
• Current parameter sets cover all channel categories except one (very
diffuse multipath), for the “office” scenarios.
• Generate additional channel realizations to cover remaining scenarios
and channel categories:
• Preferred method: find S-V parameters to fill in the holes in the
current set of realizations.
• Alternate method: supplement S-V channel realizations with
realizations taken from measured data.
Submission
Slide 22
Marcus Pendergrass
Time Domain Corporation (TDC)