Swarm Intelligence Electromagnetic Domain Countermeasure System Based on Deep Reinforcement Learning

In order to improve the capabilities of information perception, interaction and decision-making in electromagnetic domain operations, and to adapt to the development trend from single to clustered joint operations, this study proposed a swarm intelligent electromagnetic domain countermeasure system. First, it expounded on the overall framework and organizational modules of the system. The system consists of behavior interpretation and construction modules, resource scheduling and control modules, as well as functional modules such as intelligent detection, surveillance, and interference. Second, it briefly introduced the principles of intelligent algorithms. Based on reinforcement learning, it applied neural networks in deep learning as an approximation of value functions and generalized from single-agent to multi-agent. Then, this study sorted out and introduced the key technologies involved in the system, including swarm agent electromagnetic airspace and frequency domain allocation technology, swarm intelligence cognitive electronic warfare technology, cooperative allocation technology of electromagnetic resources for swarm agents and electromagnetic behavior interpretation and construction techniques. Moreover, it constructed a simple scene and explored the application of reinforcement learning in the system, and achieved good results. Lastly, the study provided an outlook on the future application scenarios of the system.