Air-to-air confrontation has attracted wide attention from artificial intelligence scholars. Since the existence of maneuvering targets and various interceptors, the mission autonomous planning of multi-target penetration strikes is difficult in unknown dynamic scenarios. A novel multi-target penetration strike autonomous planning method is proposed in this paper. First, a strike effectiveness function is designed, and dynamic target assignment is achieved through integer programming. Then, a strike planning algorithm based on deep reinforcement learning is designed to realize the autonomous decision-making of penetration in the presence of mission mutation. The simulation result shows that the method can achieve effective multi-target penetration strike in complex combat environments.
Multi-target Strike Planning in Unknown Dynamic Environment
Lect. Notes Electrical Eng.
International Conference on Autonomous Unmanned Systems ; 2022 ; Xi'an, China September 23, 2022 - September 25, 2022
Proceedings of 2022 International Conference on Autonomous Unmanned Systems (ICAUS 2022) ; Kapitel : 231 ; 2497-2509
2023-03-10
13 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Online cooperative path planning for multi-quadrotors in an unknown dynamic environment
SAGE Publications | 2022
|Path planning approach in unknown environment
British Library Online Contents | 2010
|Trajectory Planning of UAV in Unknown Dynamic Environment with Deep Reinforcement Learning
Springer Verlag | 2019
|Path Planning with Improved Dueling DQN Algorithm for UAVs in Unknown Dynamic Environment
Springer Verlag | 2024
|MULTI-TARGET DYNAMIC PLANNING METHOD FOR FLIGHT TRAJECTORY
Europäisches Patentamt | 2022
|