基于深度强化学习的多旋翼无人机空中目标自主跟踪

针对空中对接任务中的目标自主跟踪问题，提出了一种基于深度强化学习的端到端的目标跟踪方法。该方法采用近端策略优化算法，Actor网络与Critic网络共享前两层的网络参数，将无人机所拍摄图像作为卷积神经网络的输入，通过策略网络控制多旋翼无人机电机转速，实现端到端的目标跟踪，同时采用shaping方法以加速智能体训练。通过物理引擎Pybullet搭建仿真环境并进行训练验证，仿真结果表明该方法能够达到设定的目标跟踪要求且具有较好的鲁棒性。

Air Target Autonomous Tracking of Multi-rotor UAV Based on Deep Reinforcement Learning

Aiming at the problem of target autonomous tracking in the process of air docking, an end-to-end air target autonomous tracking method based on deep reinforcement learning is proposed. In this method, the near end strategy optimization algorithm is adopted. The strategy network and value network share the first two network parameters. The image captured by UAV is used as the input of convolution neural network. The motor speed of rotor UAV is controlled by strategy network to achieve end-to-end autonomous target tracking. At the same time, shaping meth-od is used to accelerate the agent training. The simulation environment is built by pybullet and the training verification is carried out. The experimental results show that the method can achieve the set target tracking requirements and has good robustness.