Research On Radar Adaptive Jamming Strategy And Evaluation Method

As artificial intelligence technology continues to flourish in various fields,the intelligent combat of targeted weapons and equipment are also developing rapidly.Multifunctional radar has become the main combat weapon of cognitive electronic warfare because of its strong adaptability and diverse anti-jamming skills.As the jamming party,it is difficult to ensure the rapid selection of the optimal jamming pattern and parameters in the fight against complex and changeable multifunctional radars.The difficulty of judging the radar working mode in real time according to the reconnaissance radar signal is further improved.This paper is based on reinforcement learning to improve the performance of jamming decision-making in complex electromagnetic environment.Intelligent jamming decision-making methods such as jamming patterns and jamming parameters for multi-function radar are studied in this paper.Aiming at the inconsistency of jamming effect evaluation caused by different modes and different jamming means,a unified evaluation index system and evaluation method of jamming effect are studied and constructed.The research contents of this paper are as follows:The jamming process of intelligent radar countermeasure system is compared and analyzed in view of the shortcomings of traditional radar jamming process,such as long time consuming,no closed-loop feedback and no dynamic intelligent adjustment.Reinforcement learning is used to model and simulate the offensive and defensive actions under the system confrontation.The jamming pattern decision method based on Q-learning is given and the influence of parameters on the algorithm is discussed.The closed-loop feedback function of observation,orientation,decision and action combat loop is considered.The evaluation index system of jamming effect based on non-cooperative parties is constructed by giving the evaluation criteria of jamming effect.A model of jamming effect evaluation based on C-support vector machine is established.In order to solve the problem of intelligent cooperative jamming decision-making against frequency agility and frequency diversity in cognitive electronic warfare,an intelligent decision-making method of parameters in frequency domain based on multi-agent hierarchical reinforcement learning is proposed.The multi-agent markov decision processes is used to construct the cooperative decision-making process of multi-jammers.A cooperative jamming decision-making in frequency domain(FD-CJDM)model is established based on the idea of hierarchical reinforcement learning.An optimization method of prioritized experience replayHierarchical Reinforcement Learning-double DQN(PER-HRL-DDQN)based on sum tree structure is adopted to find the optimal strategy.The performance of FD-CJDM based on PERHRL-DDQN is simulated,and it is verified that the proposed PER-HRL-DDQN method is obviously superior to the deep Q network method from the aspects of jamming effect and algorithm performance.According to the sequence of detected radar threat,the parameter of decision-making strategy in frequency domain can be formulated,and the design of intelligent decision-making is effectively realized in frequency domain.Aiming at the problem that the parameters of the intercepted signal are distorted due to the influence of complex electromagnetic environment in the process of interception,a discrete data processing algorithm based on improved class-attribute contingency coefficient(ICACC)standard is proposed.The correlation between the classes and features of data is maximized by introducing relevant definitions of important features and inconsistency rate.In this way,the influence of inaccurate discrimination of radar working state caused by detection error is minimized.Aiming at the problem of incomplete signal parameters caused by signal variability,a soft output basic belief assignment(BBA)correction method combined with probability similarity is adopted from the perspective of SVM output results.In the face of incomplete systems,the introduction of human errors should be reduced as much as possible while retaining valuable samples.It is necessary to improve the recognition accuracy by making full use of the known information.A jamming effect evaluation method based on the transition of radar working state is given aiming at the generation of asymmetric information caused by complex electromagnetic environment,which is the key link for the jammer to construct adaptive closedloop countermeasures.