Game-oriented radar countermeasure strategy generation method

Abstract

The invention provides a game-oriented radar countermeasure strategy generation method, which comprises the following steps: setting a countermeasure scene: setting a radar and an interference systemas two players, constructing a radar countermeasure game tree, setting two neural networks of a regret value and a countermeasure strategy and a corresponding cache region for each player, and initializing neural network parameters; traversing the K game trees within the iteration times, alternately training the radar and the interference party, training the regret value neural network by using the regret value cache region data, and then training the adversarial strategy neural network of the radar and the interference party by using the collected adversarial strategy cache region data untilthe adversarial strategy neural network converges. Compared with radar countermeasure static game research, the method has the advantages that a dynamic game model of incomplete information is constructed, Nash equilibrium of radar and interference system games is approximately solved by using a neural network under the condition that both parties are intelligent, and respective Nash equilibrium strategies are obtained through multiple times of dynamic iterative updating.

Description

technical field [0001] The invention relates to the cross-technical field of radar electronic countermeasures, game theory and artificial intelligence, in particular to a game-oriented radar countermeasure generation method. Background technique [0002] The application of artificial intelligence technology in the field of electronic countermeasures is getting more and more in-depth, and radar countermeasures are also becoming more intelligent. Both radar and jamming systems have developed intelligent algorithms with adaptive and even cognitive capabilities. Research on radar from the perspective of dynamic games between the two parties The problem of electronic countermeasures is an important development direction. [0003] At present, the field of cognitive electronic countermeasures mainly focuses on the unilateral intelligent algorithm research of radar or jammer, such as waveform optimization, target detection, identification and tracking technology of cognitive radar; ...

AI Technical Summary

Problems solved by technology

[0005] The existing radar confrontation game mainly studies the two-player zero-sum game with static complete information. Both opponents have a complete understanding of the opponent's action space and profit function, and the final confrontation strategy is also fixed. However, in the actual complex electromagnetic environment , this static strategy cannot cope with changing opponents, and it is easy for opponents to speculate on their own strategies, so that they cannot win in the confrontation

Images