MANEUVERING PENETRATION METHOD FOR MISSILE UNDER UNKNOWN INTERCEPTION STRATEGIES BASED ON TRANSFER DEEP REINFORCEMENT LEARNING
This paper is concerned with maneuvering penetration method for missile in unknown confrontation environment. Based on transfer learning method and reinforcement learning method, the penetration strategy, which includes two stages: off-line training and online confrontation training, is established. To begin with, the deep reinforcement learning method is used to realize off-line training of the missile under the typical interception strategies, in this process, the empirical interception strategies can be generated, it is referred as the off-line training process for missile. Then, together with the empirical knowledge transferred by the formula designed in this paper, the online confrontation training is carried out by reinforcement learning method when the missile is flying under unknown interception strategies, this process is online confrontation training for missile. Thus, after these two training process, the missile can select the suitable action to penetrate the interceptor successfully under unknown confrontation environment. Finally, numerical simulations are presented to verify the validity and efficiency of the method proposed.
Authors
- Juan L. G. Guirao (ORCID: https://orcid.org/0000-0003-2788-809X)
- Yaokun Wang
- Dengqing Cao (ORCID: https://orcid.org/0000-0002-3332-5704)
- Huatao Chen
- Kai Pan (ORCID: https://orcid.org/0009-0006-2871-6969)
- Kun Zhao
Institutions
- Shandong University of Technology (CN)
- Harbin Institute of Technology (CN)
- Universidad Politécnica de Cartagena (ES)
- Beijing Jingshida Electromechanical Equipment Research Institute (CN)
Publication Details
- Journal
- Fractals
- Published
- 2025-05-08
- DOI
- https://doi.org/10.1142/s0218348x25401863
- Primary Topic
- Guidance and Control Systems
- Type
- article
- Field-Weighted Citation Impact
- 0.00