DMCA
Tracking unmodified smartphones using wi-fi monitors (2012)
| Venue: | In Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems, SenSys ’12 |
| Citations: | 22 - 0 self |
BibTeX
@INPROCEEDINGS{Musa12trackingunmodified,
author = {A. B. M. Musa and Jakob Eriksson},
title = {Tracking unmodified smartphones using wi-fi monitors},
booktitle = {In Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems, SenSys ’12},
year = {2012},
pages = {281--294},
publisher = {ACM}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Smartphones with Wi-Fi enabled periodically transmit Wi-Fi messages, even when not associated to a network. In one 12-hour trial on a busy road (average daily traffic count 37,000 according to the state DOT), 7,000 unique devices were detected by a single road-side monitoring station, or about 1 device for every 5 vehicles. In this paper, we describe a system for passively tracking unmodified smartphones, based on such Wi-Fi detections. This system uses only common, off-the-shelf access point hardware to both collect and deliver detections. Thus, in addition to high detection rates, it potentially offers very low equipment and installation cost. However, the long range and sparse nature of our opportunistically collected Wi-Fi transmissions presents a significant localization challenge. We propose a trajectory estimation method based on Viterbi’s algorithm which takes second-by-second detections of a moving device as input, and produces the most likely spatio-temporal path taken. In addition, we present several methods that prompt passing devices to send additional messages, increasing detection rates an use signal-strength for improved accuracy. Based on our experimental evaluation from one 9-month deployment and several single-day deployments, passive Wi-Fi tracking detects a large fraction of passing smartphones, and produces high-accuracy trajectory estimates. 1
Keyphrases
unmodified smartphones wi-fi monitor additional message likely spatio-temporal path average daily traffic count unique device experimental evaluation wi-fi transmission sparse nature use signal-strength improved accuracy several single-day deployment high detection rate significant localization challenge off-the-shelf access point hardware state dot second-by-second detection 12-hour trial present several method busy road large fraction installation cost long range high-accuracy trajectory estimate single road-side monitoring station low equipment passive wi-fi tracking transmit wi-fi message wi-fi detection trajectory estimation method viterbi algorithm 9-month deployment detection rate
