Research On Distributed Multi-Agent Cooperative Jamming And Evaluation Methods In Cognitive Warfare

Information warfare is beginning to appear on the stage of human warfare.Iin the face of the highly dynamic and real-time electromagnetic environment in the information battlefield,how to make use of the resource-limited UAV clusters to jam against radars is an important research direction.This thesis focuses on three aspects of the online resource allocation problem for jammer clusters,the collaborative decision-making method for multiple jammers,and jamming evaluation,in order to make full use of the limited jamming resources,flexibly generate jamming waveforms applicable to complex radar systems,and improve the effectiveness of jamming against the enemy.The main work of this paper is divided into three parts as follows.Aiming at the problem of resource constrained UAV clusters in complex electromagnetic environments,this thesis is constrained to establish a jamming resource allocation optimization model,determine the optimization objectives of resource allocation,study the relative relationship between jamming cost effectiveness ratio and jamming resources and jamming benefits,followed by determining the resource allocation cost effectiveness ratio based on the threat level and stability of jamming in different operating states of radars,designing and implementing a multi-jammer simulation A multi-jammer simulation system is designed and implemented,and the stability of the online system is introduced to further regulate the jamming resource allocation.The simulation experiments demonstrate that the resource allocation model proposed in this thesis can effectively integrate resources and improve the efficiency of the jamming system.In order to reduce the dimension of the decision space and improve the speed of the single jamming decision generation,this thesis establishes a state transfer table of jamming patterns based on the working characteristics of phased array radars to select the appropriate jamming pattern when the radar working state changes.The overall information of the cluster is used to update its own state transfer table with the reception and perception parameters of neighboring UAVs,thus accelerating the iteration rate and enabling it to reach the convergence state faster.Simulation experiments demonstrate that the multi-jammer collaborative decision model proposed in this thesis can learn the radar mode switching and parameter selection strategies,and the speed of convergence of the jamming model to the global optimal solution is increased.In order to address the problem of effectiveness evaluation based on subjective perception parameters on the jamming side,this thesis proposes a jamming effectiveness evaluation index system based on the radar operating parameters obtained by the jammers,and reduces the redundancy between the indexes by the great irrelevance method,and calculates the contribution degree of each jammer in a confrontation,which improves the practicality of jamming effectiveness evaluation.Simulation experiments demonstrate that the method can effectively assess jamming effectiveness and provide accurate feedback information for the adaptive jamming decision unit,which is helpful for jamming resource allocation and strategy optimization.