Motion sensing using WLAN signal fluctuations

K. Kavitha Muthukrishnan, M.E.M. Lijding, Nirvana Meratnia, Paul J.M. Havinga

Research output: Book/ReportReportProfessional

Abstract

The ability to infer the motion of the user has previously been possible only with the usage of additional hardware. In this paper we show how motion sensing can be obtained just by observing the WLAN radio’s signal strength and its fluctuations. For the first time, we have analyzed the signal strength fluctuations in three different domains: (i) time, (ii) frequency, and (iii) space. Our analysis in all these three domains confirmed our claim that ��?when a device is moving, signal strengths from all heard access points vary much greater and more obvious compared to when a device is still��?. Using this observation, we present two algorithms Frequency-Spread Motion Detection and Spatially-Spread Motion Detection to infer the motion of the user. A two-state classification scheme is used in both algorithms to deduce the user motion as either ’still’ or ’moving’. The results and performances are benchmarked against the ground truth using a machine learning toolkit. Both these algorithms show an overall classification accuracy of 95% and 97% respectively, and use only the radio signal that the wireless device should have to perform its normal operation. Finally we discuss how adding motion status inferred from our motion detection algorithms at runtime improves accuracy of our calibration-free localization algorithm to <5m.
Original languageUndefined
Place of PublicationEnschede
PublisherCentrum voor Telematica en Informatie Technologie
Number of pages20
Publication statusPublished - 21 Dec 2006

Publication series

NameCTIT Technical Report Series
PublisherCentre for Telematics and Information Technology, University of Twente
No.06-76
ISSN (Print)1381-3625

Keywords

  • METIS-237418
  • EWI-8163
  • CAES-PS: Pervasive Systems
  • IR-63691

Cite this

Kavitha Muthukrishnan, K., Lijding, M. E. M., Meratnia, N., & Havinga, P. J. M. (2006). Motion sensing using WLAN signal fluctuations. (CTIT Technical Report Series; No. 06-76). Enschede: Centrum voor Telematica en Informatie Technologie.
Kavitha Muthukrishnan, K. ; Lijding, M.E.M. ; Meratnia, Nirvana ; Havinga, Paul J.M. / Motion sensing using WLAN signal fluctuations. Enschede : Centrum voor Telematica en Informatie Technologie, 2006. 20 p. (CTIT Technical Report Series; 06-76).
@book{a2fbb237d5414c7db69e3a2be0fc7b59,
title = "Motion sensing using WLAN signal fluctuations",
abstract = "The ability to infer the motion of the user has previously been possible only with the usage of additional hardware. In this paper we show how motion sensing can be obtained just by observing the WLAN radio’s signal strength and its fluctuations. For the first time, we have analyzed the signal strength fluctuations in three different domains: (i) time, (ii) frequency, and (iii) space. Our analysis in all these three domains confirmed our claim that ��?when a device is moving, signal strengths from all heard access points vary much greater and more obvious compared to when a device is still��?. Using this observation, we present two algorithms Frequency-Spread Motion Detection and Spatially-Spread Motion Detection to infer the motion of the user. A two-state classification scheme is used in both algorithms to deduce the user motion as either ’still’ or ’moving’. The results and performances are benchmarked against the ground truth using a machine learning toolkit. Both these algorithms show an overall classification accuracy of 95{\%} and 97{\%} respectively, and use only the radio signal that the wireless device should have to perform its normal operation. Finally we discuss how adding motion status inferred from our motion detection algorithms at runtime improves accuracy of our calibration-free localization algorithm to <5m.",
keywords = "METIS-237418, EWI-8163, CAES-PS: Pervasive Systems, IR-63691",
author = "{Kavitha Muthukrishnan}, K. and M.E.M. Lijding and Nirvana Meratnia and Havinga, {Paul J.M.}",
year = "2006",
month = "12",
day = "21",
language = "Undefined",
series = "CTIT Technical Report Series",
publisher = "Centrum voor Telematica en Informatie Technologie",
number = "06-76",

}

Kavitha Muthukrishnan, K, Lijding, MEM, Meratnia, N & Havinga, PJM 2006, Motion sensing using WLAN signal fluctuations. CTIT Technical Report Series, no. 06-76, Centrum voor Telematica en Informatie Technologie, Enschede.

Motion sensing using WLAN signal fluctuations. / Kavitha Muthukrishnan, K.; Lijding, M.E.M.; Meratnia, Nirvana; Havinga, Paul J.M.

Enschede : Centrum voor Telematica en Informatie Technologie, 2006. 20 p. (CTIT Technical Report Series; No. 06-76).

Research output: Book/ReportReportProfessional

TY - BOOK

T1 - Motion sensing using WLAN signal fluctuations

AU - Kavitha Muthukrishnan, K.

AU - Lijding, M.E.M.

AU - Meratnia, Nirvana

AU - Havinga, Paul J.M.

PY - 2006/12/21

Y1 - 2006/12/21

N2 - The ability to infer the motion of the user has previously been possible only with the usage of additional hardware. In this paper we show how motion sensing can be obtained just by observing the WLAN radio’s signal strength and its fluctuations. For the first time, we have analyzed the signal strength fluctuations in three different domains: (i) time, (ii) frequency, and (iii) space. Our analysis in all these three domains confirmed our claim that ��?when a device is moving, signal strengths from all heard access points vary much greater and more obvious compared to when a device is still��?. Using this observation, we present two algorithms Frequency-Spread Motion Detection and Spatially-Spread Motion Detection to infer the motion of the user. A two-state classification scheme is used in both algorithms to deduce the user motion as either ’still’ or ’moving’. The results and performances are benchmarked against the ground truth using a machine learning toolkit. Both these algorithms show an overall classification accuracy of 95% and 97% respectively, and use only the radio signal that the wireless device should have to perform its normal operation. Finally we discuss how adding motion status inferred from our motion detection algorithms at runtime improves accuracy of our calibration-free localization algorithm to <5m.

AB - The ability to infer the motion of the user has previously been possible only with the usage of additional hardware. In this paper we show how motion sensing can be obtained just by observing the WLAN radio’s signal strength and its fluctuations. For the first time, we have analyzed the signal strength fluctuations in three different domains: (i) time, (ii) frequency, and (iii) space. Our analysis in all these three domains confirmed our claim that ��?when a device is moving, signal strengths from all heard access points vary much greater and more obvious compared to when a device is still��?. Using this observation, we present two algorithms Frequency-Spread Motion Detection and Spatially-Spread Motion Detection to infer the motion of the user. A two-state classification scheme is used in both algorithms to deduce the user motion as either ’still’ or ’moving’. The results and performances are benchmarked against the ground truth using a machine learning toolkit. Both these algorithms show an overall classification accuracy of 95% and 97% respectively, and use only the radio signal that the wireless device should have to perform its normal operation. Finally we discuss how adding motion status inferred from our motion detection algorithms at runtime improves accuracy of our calibration-free localization algorithm to <5m.

KW - METIS-237418

KW - EWI-8163

KW - CAES-PS: Pervasive Systems

KW - IR-63691

M3 - Report

T3 - CTIT Technical Report Series

BT - Motion sensing using WLAN signal fluctuations

PB - Centrum voor Telematica en Informatie Technologie

CY - Enschede

ER -

Kavitha Muthukrishnan K, Lijding MEM, Meratnia N, Havinga PJM. Motion sensing using WLAN signal fluctuations. Enschede: Centrum voor Telematica en Informatie Technologie, 2006. 20 p. (CTIT Technical Report Series; 06-76).