A UAV Swarm Target Defense Method Based on Deep Reinforcement Learning

Abstract

The present invention provides a UAV cluster target defense method based on deep reinforcement learning. In the training stage, the initial position information and behavior information of the incoming target are obtained, and a deep reinforcement learning training neural network is established, and the neural network is trained. In the execution stage, the The UAV cluster status and target status are input to the trained deep neural network, and the output results are judged. The invention expands the state space and behavior space of the UAV cluster task, constructs a unified decision-making network for the task-oriented UAV cluster, and realizes the unified command and control of the decision-making network for an indefinite number of UAVs.

Description

technical field [0001] The invention relates to the fields of machine learning, path planning and target confrontation, in particular to an unmanned swarm target defense method. Background technique [0002] Nowadays, the research on UAV swarms has attracted more and more attention from everyone, and has made certain progress in the direction of UAV swarm mission decision-making, information interaction and information fusion between UAVs, collaborative detection, path planning, and interaction methods. As a result of the research, among the mission sequences of many UAV swarms, the defense mission is the most important thing that cannot be bypassed. In the battlefield environment, it is often necessary to defend one's own important strategic goals and prevent enemy targets from attacking. Therefore, deploying drones around the base to prevent possible enemy attacks will become a drone cluster. one of the important tasks. [0003] The current research on the behavior contr...

AI Technical Summary

Problems solved by technology

Judging from the achievements of the coordinated missions of UAV clusters realized in China, my country's UAV clusters have more research on how to effectively prevent collisions when UAV clusters perform tasks collaboratively, and the payload of UAVs and its execution are more complex. High-level tasks, such as: coordinated detection, coordinated attack, coordinated defense and other task-level related research are still relatively few

Images