TY - JOUR
T1 - Robust WLAN-Based Indoor Intrusion Detection Using PHY Layer Information
AU - Lv, Jiguang
AU - Man, Dapeng
AU - Yang, Wu
AU - Du, Xiaojiang
AU - Yu, Miao
N1 - Publisher Copyright:
© 2013 IEEE.
PY - 2017/12/19
Y1 - 2017/12/19
N2 - Intrusion detection techniques are widely used to guarantee the security of people's possessions. With the rapid development of wireless communication, device-free passive human detection based on wireless techniques may have more opportunities in intrusion detection. WiFi has been widely deployed in both public and private areas, which can be used as generalized sensors to detect human motion beyond communication. As a result, there have been several researches on WLAN-based motion detection. However, the detection accuracy of previous approaches declines significantly when people's moving speed becomes very slow. In this paper, we explore a novel method which has a relative stable detection performance under different moving speeds. We extract a novel feature representing the fluctuation of the whole channel from channel state information at the physical layer of 802.11n wireless networks, and utilize a probability technique to detect human motion. A hidden Markov model is leveraged as the classifier to make human detection a probability problem. We implement the system using off-the-shelf WiFi devices and evaluate it in two scenarios. As indicated in the evaluation results, our approach is an appropriate method for intrusion detection.
AB - Intrusion detection techniques are widely used to guarantee the security of people's possessions. With the rapid development of wireless communication, device-free passive human detection based on wireless techniques may have more opportunities in intrusion detection. WiFi has been widely deployed in both public and private areas, which can be used as generalized sensors to detect human motion beyond communication. As a result, there have been several researches on WLAN-based motion detection. However, the detection accuracy of previous approaches declines significantly when people's moving speed becomes very slow. In this paper, we explore a novel method which has a relative stable detection performance under different moving speeds. We extract a novel feature representing the fluctuation of the whole channel from channel state information at the physical layer of 802.11n wireless networks, and utilize a probability technique to detect human motion. A hidden Markov model is leveraged as the classifier to make human detection a probability problem. We implement the system using off-the-shelf WiFi devices and evaluate it in two scenarios. As indicated in the evaluation results, our approach is an appropriate method for intrusion detection.
KW - Device-free passive
KW - channel state information
KW - dynamic speed
KW - intrusion detection
UR - http://www.scopus.com/inward/record.url?scp=85039764001&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85039764001&partnerID=8YFLogxK
U2 - 10.1109/ACCESS.2017.2785444
DO - 10.1109/ACCESS.2017.2785444
M3 - Article
AN - SCOPUS:85039764001
VL - 6
SP - 30117
EP - 30127
JO - IEEE Access
JF - IEEE Access
ER -