Optimal multi-channel transmission energy allocation against jamming attacks via data-driven reinforcement learning under Stackelberg game framework

Employing a Markov Stackelberg game and Signal-to-Interference-plus-Noise Ratio (SINR) model, this research develops a strategic framework for sensor power scheduling in Cyber-Physical Systems (CPSs) under adversarial conditions. It innovates by determining energy-efficient sensor transmission strategies against jamming, using data-driven reinforcement learning to optimize power allocation adaptively. The necessary training parameter design methodology is presented by the study, and stability proofs are provided. Simulation results demonstrate that the proposed algorithm outperforms existing approaches in reducing estimation errors and mitigating the attacker’s impact.