Spoofing detection games in underwater sensor networks

Spoofing detection is essential in underwater sensor networks (UWSNs), especially for safety-critical applications. Because of the high bit error rates, low bandwidth and large delay of the acoustic channels, it is challenging to detect spoofing attacks in UWSNs. In this paper, we propose spoofing detection strategies based on reinforcement learning. The interactions between a surface station and an underwater spoofer are formulated as a zero-sum game. The Nash equilibrium (NE) of the spoofing detection game in UWSNs is derived and the existing condition for a unique NE is provided. A spoofing detection method based on Q-learning is proposed for dynamic underwater sensor networks. Simulation results have validated the efficacy of the proposed spoofing detection scheme in UWSNs.