As the use of mobile, wireless computing devices continues to grow, so does concern with assuring that communications within these wireless networks is not vulnerable to various kinds of attacks. Physical-layer Network Coding (PNC) was developed as a means for improving communication in wireless networks. This project will conduct exploratory research to determine whether and how the use of PNC can be exploited to detect attacks. The work is organized into three tasks. The first task will identify capabilities to detect attacks and design corresponding mechanisms for differing network infrastructures and attacker models. Signal strength attacks, Sybil attacks, and wormhole attacks will be the initial targets of study. The second task will study the detection accuracy of the proposed mechanisms and explore more complex network scenarios. The third task will develop simulations to model the attack detection mechanisms.
The overall research goal of this project is as follows: we seek to explore and deeply understand the security capabilities of the physical layer network coding (PNC) technique in wireless networks, build new attack detection mechanisms upon the gained insights of PNC, and evaluate the approaches through both theoretical analysis and simulation. In wireless networks, bandwidth is considered one of the most scarce and valuable resources. The PNC technique has been proposed to greatly improve the bandwidth usage efficiency. However, very little research has been conducted to study its security capabilities. In this project, we investigate the problem from a new perspective and create new solutions that are different from existing approaches. Instead of trying to protect network coding results from pollution attacks, we seek to explore whether or not the PNC technique can be used to detect attacks on wireless networks. Specifically, our research: (1) identifies the properties of PNC that can be used for attack detection; creates the design of the detection mechanisms under different network structures and attacker models; (2) identifies and overcomes the difficulties of turning the design into practical solutions in real networks; and (3) evaluates the approaches through both theoretical analysis and simulation. The proposed research leads to the design of innovative detection mechanisms of multiple types of malicious attacks on wireless networks including: Sybil attacks, primary user emulation attacks, wormhole attacks, and sender authentication. The proposed research can also be used to achieve accurate localization in wireless networks. From student training side, three PhD students and one Master’s student benefit from the project in thesis topic selection, innovative research, problem solving, and academic writing. The students are rooted in either EE or CS major which shows the interdisciplinary nature of the project. These students published three journal papers, three conference papers, and one poster during the project duration. From the education side, the PI designed new educational modules on physical layer security in wireless networks. These modules have been integrated into both graduate level and undergraduate level security courses. More than ten presentations and invited talks have been given by the research team to promote and disseminate the research achievements. The general public could also benefit from the research achievements of the project. Wide adoption of the PNC technique will help address the bandwidth deficiency issue and enable the development and deployment of new bandwidth-hungry mobile applications in the fields of entertainment, public safety, and interactive education. Therefore, a large number of end users could benefit from our research results.
