【摘要】 本文针对固定翼无人机自主着陆控制问题,提出了基于深度强化学习(DRL)的无人机着陆轨迹跟踪控制方法。首先,搭建了小型固定翼无人机Ultra Stick 25E的仿真模型,设计了满足过程和终端约束的着陆参考轨迹。其次,提出了基于深度确定性策略梯度(DDPG)的无人机一体化控制框架,设计了考虑跟踪误差和轨迹平稳性的奖励函数。最后,通过离线训练,得到了轨迹跟踪一体化控制器。仿真试验结果表明,本文提出的方法比传统PID控制方法精度更高。更多还原

【Abstract】 Focusing on the problem of autonomous landing control of fixed-wing UAVs, this paper proposes a tracking control method for UAV landing trajectory based on Deep Reinforcement Learning(DRL). First, we built a simulation model of the small fixed-wing UAV Ultra Stick 25E and designed a landing reference trajectory that satisfies the process and terminal constraints. Second, we proposed a UAV-integrated control framework based on Deep Deterministic Policy Gradient(DDPG) and designed a reward function considering tracking error and trajectory stability. Finally, through the offline training, we obtained the trajectory tracking integrated controller. The simulation results show that the proposed method is more accurate than the traditional PID control method.更多还原