Monte Carlo Counterfactual Regret Minimization for Waveform Optimization in Radar and Jammer Games

With the continuous development of high-speed signal processing hardware and various machine learning al-gorithms, the adaptive ability of radar and jammer has been significantly improved, leading to an escalating confrontation in the electromagnetic field. This paper studies the waveform domain game between cognitive MIMO radar and intelligent jammer. Considering the presence of imperfect information during the interaction between radar and jammer, this paper uses an imperfect information extensive-form game model to model the problem. The Minimum Mean Square Error (MMSE) of the radar's target parameter estimation capability is used as the game utility function. The Monte Carlo Counterfactual Regret Minimization (MCCFR) algorithm is employed to obtain the optimal waveforms for both radar and jammer when the game reaches an approximate Nash Equilibrium. Simulations verify the effectiveness of this algorithm, demonstrating its ability to converge to the optimal strategy. This provides compelling support for swift optimization decision-making in the waveform level, a crucial aspect of future electronic warfare.