6G ISAC security: Game-theoretic RL detects beamforming attackers

Researchers Parmida Geranmayeh and Onur Günlü have published a paper on arXiv (2607.06115) addressing a critical security challenge in next-generation 6G networks: detecting attackers who manipulate beamforming in integrated sensing and communication (ISAC) systems. In ISAC, the same wireless infrastructure handles both data transmission and environmental sensing, enabling systems to monitor user behavior and channel variations. Attackers can exploit this by deliberately altering beamforming directions to increase interference, tricking the transmitter into directing its main lobe toward the attacker instead of legitimate users.

The proposed solution combines game-theoretic modeling with reinforcement learning. Legitimate users and the attacker are modeled as players in a game, with utility functions capturing objectives like minimizing interference or maximizing signal quality. These utility-based formulations are then fed into an RL framework that learns optimal beamforming strategies while simultaneously detecting malicious behavior. Simulation results demonstrate that the distributed game-theoretic RL approach effectively identifies and counters beamforming attacks in dynamic urban 6G scenarios, offering a new layer of security for future wireless networks.