A Sophisticated Anti-Eavesdropping Strategy

Abstract

Wireless networks are susceptible to malicious attacks, especially those involving eavesdropping. In this letter, we consider a new type of anti-eavesdropping strategy which, beyond the basic goal of increasing the secrecy rate, also wants to achieve this in the most unpredictable way for the adversary. We model the problem by a non-zero sum game where a control center (called the transmitter) must communicate with a group of nodes allocated in security zone in the presence of an adversary intent on eavesdropping upon this communication. The transmitter wants to find a trade-off between two goals: (a) to increase the expected secrecy rate, and (b) to maintain such secret communication in the most unpredictable way for the adversary. As a metric for unpredictability of the transmitter we consider the Shannon entropy of its strategy. We model this problem by a nonzero-sum, two-player resource allocation game. The equilibrium is found in closed form, and its dependence on communication network parameters is illustrated. Finally, weighting coefficients for the basic and secondary goals of the transmitter are optimized based on proportional fairness criteria.