Anti-jamming Equilibrium Strategy Learning of Frequency Agile Radar based on Monte Carlo Tree Search

Frequency agility is a promising technique for addressing the threat of mainlobe jamming in modern electronic warfare. This paper investigates the confrontation between a subpulse-level frequency agile radar and a smart mainlobe jammer from a game theory perspective. The interaction process between the two sides is modeled as a two-player zero-sum extensive form game, which can capture the multi-stage and highly competitive nature. To obtain the optimal radar frequency agility strategy in the presence of a smart jammer, we introduce the Simultaneous Move Monte Carlo Tree Search (SM-MCTS) algorithm, which improves upon the classic Monte Carlo Tree Search algorithm in both selection and backpropagation steps. Furthermore, we discuss three selection functions that choose actions for the radar and jammer at each game tree node. Finally, simulation results demonstrate the effectiveness of the SM-MCTS algorithm for approximating the Nash equilibrium.