Knowledge Based Communication Jamming Decisions

Communication jamming decision-making is a key technical difficulty in the research of cognitive electronic warfare.In the complex electromagnetic environment,the tradi-tional communication interference decision algorithm has some problems,such as slow convergence and poor interference effect,when aiming at the unknown working state of the enemy,which can not meet the flexibility and real-time requirements of modern cog-nitive electronic warfare.Aiming at the above problems,based on the knowledge-driven idea,this thesis mines and models the knowledge hidden in the communication jamming historical data,and designs a knowledge-based communication jamming decision-making algorithm.There are two innovations in this thesis:Firstly,aiming at the unknown communication mode state in the ”point-to-point”communication scenario,this thesis proposes a knowledge-based jamming action recom-mendation algorithm,which solves the problem of slow convergence of the traditional interference decision algorithm based on reinforcement learning.The algorithm models the jamming action of the interferer and the communication state of the victim as entities in the interference knowledge graph,and extracts the relationship between entities based on the correlation of historical data and attributes.This algorithm uses the interaction data between the unknown communication mode state of the interferer and the victim and the relationship between entities to evaluate the similarity between the unknown com-munication mode state of the victim and the communication mode state recorded in the interference history database,and enables the recommendation of the best interference ac-tion based on the similarity.The simulation experiment results show that,compared with the traditional reinforcement learning-based interference decision-making algorithm,the knowledge-based jamming action recommendation algorithm can achieve better jamming effect in a shorter iteration period,and the average bit error rate is higher than the tradi-tional algorithm out about 83%Secondly,aiming at the situation that the topology structure is unknown in the ”many-to-many” communication networking scenario,this thesis proposes a knowledge-based bayesian neural network interference decision-making algorithm,which solves the prob-lems of slow convergence and poor interference effect of the traditional algorithm.By mining the importance of communication networking nodes hidden in the interference historical data,the algorithm guides the jammer to select the optimal combination of com-munication networking nodes for interference and can update the knowledge through the increasing interference historical data,so as to continuously optimize the interference de-cision.The simulation results show that the knowledge-based bayesian neural network interference decision-making algorithm has better generalization in different types of com-munication networking scenarios compared with other interference decision-making algo-rithms that also do not rely on prior information of the communication network topology,as well as better interference effect.The cumulative number of truncated flows is about6% higher than other interference decision-making algorithms that do not rely on prior information.This thesis studies the problem of communication jamming decision-making in the context of cognitive electronic warfare based on the thoughts of knowledge-driven,fully excavates and properly models the knowledge in jamming historical data,so that it can assist jammers in jamming decision-making and help the latter to have better real-time and flexibility in complex electromagnetic environment.