Design of Cognitive Jamming Decision-Making System Against MFR Based on Reinforcement Learning

Abstract

Electronic countermeasures are developing towards intelligence. The multifunctional radar changes its working state in real time according to the task requirements. The traditional jamming decision-making method can not quickly adjust the jamming mode according to the jamming effect and environmental changes. It is not suitable for complex and changeable multifunctional radar. For this problem, a cognitive jamming decision-making system based on reinforcement learning is designed. For the evaluation of jamming effect, an evaluation method based on Improved Sparrow Search Algorithm-Support Vector Machine (ISSA-SVM) is proposed, which can evaluate the jamming effect online. The results are fed back to the jammer to provide basis for jamming decision-making. For the jamming decision-making process, the interference experience table is combined with Heuristic Accelerated Q-Learning (HAQL). A jamming decision-making method based on adaptive HAQL algorithm is proposed, which adaptively adjusts the jamming mode and jamming power according to the change of radar threat level. A one-to-one interference scenario is established and simulated. The results show that the system can realize the closed-loop cognitive interference of learning and confrontation.