Lei Wang Ke Sun Haipeng Dai Alex X Liu Xiaoyu Wang Nanjing University Michigan State University SECON18 June 13 th 2018 2 22 Gesture tracking inspires various applications ID: 687389
Download Presentation The PPT/PDF document "WiTrace : Centimeter-Level Passive Ges..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
WiTrace: Centimeter-Level Passive Gesture Tracking Using WiFi Signals
Lei Wang*, Ke Sun*, Haipeng Dai*, Alex X. Liu*^, Xiaoyu Wang*
Nanjing University*, Michigan State University^
SECON'18 June 13
th
, 2018Slide2
2
/22Gesture tracking inspires various applications.Tracking with WiFi is superiorUbiquitous:
almost everywhere.Non-invasive: not wearing/carrying any devices and protect privacy.Not limited: lighting condition or room layout.
AR assistance
VR assistance
Selecting menu
Playing game
MotivationSlide3
Google project soli
Mobicom'16 LLAP
NSDI
’
14,
WiTrack
NSDI
’
15,WiTrack2.0
3
/22
MobiHoc'17 Widar
FMCW signal need a
high bandwidth
of 1.79 GHz !
Mobicom'15 WiKey
UbiComp'16 WiFinger
INFOCOM '15 WiGest
Though using Wi-Fi, these solutions focus on
training-based activity recognition,
yet not tracking.
They have either
limited range
or
tracking resolution
!
MotivationSlide4
Can we build a gesture tracking system: Using WiFi signals?
With high precisionWith large working range4/22
Problem StatementSlide5
Challenge-1: What characteristics of WiFi can be leveraged to achieve cm-level tracking precision?
Solution: CSI phaseAdvantage: CSI provides more information than other WiFi characteristics (RSSI). CSI Phase has higher precision over CSI amplitude.
5/22
CSI Phase ModelSlide6
CSI Phase Model
Illustration of multiple paths
6/22Slide7
1-D TrackingDenoise the CSI signalHampel filter
Average moving filterDetect the movement
7/22Slide8
1-D TrackingChallenge-2:
How to seperate the phase changes caused by moving hands from CSI values due to other environments?Existing work: DDBR: low surrounding noise and can hardly detect slow movement. LEVD: difficult to reliably detect the local maximum and minimum points
8/22Slide9
1-D TrackingExtracting Static Component (ESC):
Find alternate maximum and minimum points that are lzrger than the emperical threshold.STFT to derive the instaneous Doppler frequency shift.Remove extreme points smaller than threshold.Average adjacent two points to derive the static value.
9/22Slide10
1-D TrackingESC vs. LEVD:ESC improves the
robustness to small ambinent noise than LEVDESC is more sensitive to small body movement than LEVD
ESC vs LEVD
I/Q trace of
raw CSI
I/Q trace of
dynamic vector
10
/22Slide11
2-D TrackingChallenge-3: H
ow to estimate the initial position of hand in 2-D space?Existing work: mTrack: discrete beam scanning mechanism to pinpoint the object's initial localization. LLAP: IDFT to process CFR signals for all subcarriers to estimate the absolute position.Basic idea: Two preamble gestures to measure the initial position of hand.
11/22Slide12
2-D TrackingInitial Position EstimationUser push hand along
x-axis and y-axis;Set the grid as the candidate initial position;Calculate the tracking trajectory for two receivers based on the initial position and path change for two directions.
12/22Slide13
2-D TrackingInitial Position Estimation
Find N candidate positions which have the N top smallest deviations and for x-axis and y-axis, respectively.Calculate N*N distance matrix , where Find the smallest element in the matrix and average the coordinate value.
13/22Slide14
2-D TrackingInitial Position Estimation
14/22Slide15
2-D TrackingSuccessive 2-D trackingEstimate
the initial hand positionSolve two equations corresponding to two receiversTrajectory CorrectionKalman filter based on CWPA model
15/22Slide16
ImplementationDevices
3 USRP-N2102 links (1 per receiver)Parameters:20 MHz bandwith64 CSI subcarriersCentral frequency at 2.4GHzTx power: 20dBm
1D scenario
2
D scenario
16
/22Slide17
Experiment1-D tracking performanceWiTrace achieves average error of
1.46 cm and 4.99 cm with and without the plank. WiTrace achieves average error of 3.75 cm and 2.51 cm for omnidirectional antenna and directional antenna.ESC achieves better performance than other algorithms.
17/22Slide18
Experiment1-D tracking performanceWiTrace is
robust to background activities which are 2 m away from the receiver for different users.WiTrace achieves average tracking error of 6.46 cm and 3.80 cm while pushing hand at different heights and walking around, respectively.
18/22Slide19
Experiment2-D tracking performanceWiTrace achieves average
3.91 cm estimated error with the template,and average 10.18 cm error without template for intial position estimation.
19/22Slide20
Experiment2-D tracking performanceWiTrace achieves an average tracking error of
2.09 cm for three shapes' trajectory (i.e., rectangle, triangle, and circle).20/22Slide21
ConclusionsWiTrace achieves
high accuracy gesture tracking using WiFi signals.We propose a novel scheme based on two preamble gestures to measure the initial position of hand.We implement WiTrace on USRP.
21/22Slide22
Q&A
22/22