Securing Wireless Localization against Signal Strength Attacks
报告人:
Yingying Chen
Prof. of Stevens Institute of Technology
报告时间:
2008-06-19 09:00
报告地点:
东主楼10－103
主办单位:
计算机系
简介:
Obtaining accurate positions of nodes in wireless and sensor networks is important because the location of wireless devices is a critical input to many high-level services. Such services include healthcare monitoring, wildlife animal habitat tracking, emergency rescue and recovery, location-based access control, and location-aware content delivery. However, the localization infrastructure can be subjected to non-cryptographic attacks which cannot be addressed by traditional security services. Further, location services are only useful if the location information is accurate and trustworthy. Thus it is desirable to explore solutions to detect and eliminate these attacks from the network. In this talk, I first describe a generalized localization model and a representative set of localization algorithms. Then, I present our study on the robustness of these algorithms to signal strength attacks.
Next, I propose several attack detection schemes for wireless localization systems. We formulate a theoretical foundation for the attack detection problem using statistical significance testing. Our experimental results provide strong evidence of the effectiveness of our approaches with high detection rates and low false positive rates across both an 802.11 (WiFi)) network as well as an 802.15.4 (ZigBee) network in two real office buildings. Further, we propose a scheme using K-means clustering analysis for both detecting identity-based spoofing attacks as well as localizing the positions of the adversaries so that to eliminate the attacks from the network. Additionally, we present the GRAIL (General Real-Time Adaptable Indoor Localization) system which is a practical implementation of wireless localization. It is designed as a distributed system with multiple software modules that can localize a wide range of wireless devices in challenging environments using customized statistical Bayesian Networks.
