一种基于强化学习的UAV目标搜索算法*

针对未知环境中无人机（unmanned aerial vehicles, UAV）目标搜索问题进行研究。建立UAV目标搜索模型，将强化学习理论应用于目标搜索问题中。提出一种未知环境中基于Q学习的UAV目标搜索算法，并将其与基于D-S证据理论的UAV搜索方法进行仿真比较。仿真结果显示此算法收敛且UAV快速搜索到了目标，此结果表明，通过对UAV在设定条件下的强化学习训练，可以使其具备一定的环境适应能力，UAV在没有任何目标信息的战场环境中能够有效执行搜索任务。

Algorithm based on reinforcement learning for UAV search

This paper stadied the problem of target search for an unknown area by UAV.It used reinforcement learning in search problem considering the UAV's search model.And proposed the algorithm of Q-learning search in an unknown surrounding and of which compared the simulations results with D-S method.The result shows that the UAV can adapt to the environment after training in reinforcement learning,and execute the search mission without any information about the targets testifying the algorithm's validity.