Scholarlink

The design of guidance laws for launch vehicles in the gale area necessitates a delicate balance between achieving high control accuracy, ensuring robustness against significant disturbances, and minimizing overload to safe guard against potential damage to the launch vehicle’s structure. In this paper, we develop network formed adaptive guidance laws for various flight tasks, which are trained with a reinforcement learning algorithm to address the challenges mention above. The simulation results demonstrate that the guidance laws trained with our algorithm exhibit robust characteristics under the influence of various wind disturbances and accurately track the desired flight trajectory with high precision throughout the flight process. Furthermore, in the gale area, the guidance laws enable the launch vehicle to operate with reduced overload while maintaining an acceptable level of tracking accuracy. This algorithm represents a promising candidate for designing guidance laws for launch vehicles in the gale area or under unknown disturbances during the flight process.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Robust Guidance Law Design of Launch Vehicles via Deep Reinforcement Learning Algorithm

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Chenhui Jia,
  • Xiaodong Liu,
  • Xiaomin Li,
  • Qinghai Gong,
  • Zhaolei Wang,
  • Ran Li

摘要

The design of guidance laws for launch vehicles in the gale area necessitates a delicate balance between achieving high control accuracy, ensuring robustness against significant disturbances, and minimizing overload to safe guard against potential damage to the launch vehicle’s structure. In this paper, we develop network formed adaptive guidance laws for various flight tasks, which are trained with a reinforcement learning algorithm to address the challenges mention above. The simulation results demonstrate that the guidance laws trained with our algorithm exhibit robust characteristics under the influence of various wind disturbances and accurately track the desired flight trajectory with high precision throughout the flight process. Furthermore, in the gale area, the guidance laws enable the launch vehicle to operate with reduced overload while maintaining an acceptable level of tracking accuracy. This algorithm represents a promising candidate for designing guidance laws for launch vehicles in the gale area or under unknown disturbances during the flight process.