This paper considers using a team of solar-powered UAVs to collaboratively eavesdrop on a mobile ground target in an urban environment. A practical application is that the police department uses UAVs to collect the transmitted data from a criminal suspect so that the transmitted content can be understood. Using UAVs in this task is a much more cost-effective and also less suspicious option than the conventional surveillance by a police helicopter. We focus on the online and decentralized path planning for the solar-powered UAV team. We formulate a multi-objective optimization problem which forces the UAVs to complete the uninterrupted eavesdropping task, not enter any No-fly space (NFS), keep a safe distance from each other and harvest as much energy as possible. We propose a Rapidly-exploring Random Tree (RRT) based path planning method, consisting of an initial planning phase and an online planning phase. In the online planning phase, each UAV plans its own path based on the shared information across the team. Since the UAV model can be taken into account in the planning phase, the obtained UAV paths are feasible without any later adjustment. More importantly, this method is easily implementable in real-time.

本文考虑使用一支太阳能无人飞行器团队，共同侦听城市环境中的移动地面目标。实际应用是警察部门使用无人机从犯罪嫌疑人那里收集传输的数据，以便可以理解传输的内容。与传统的警用直升机监视相比，在这种任务中使用无人机是一种更具成本效益的方法，而且可疑程度也较小。我们专注于太阳能无人机团队的在线和分散式路径规划。我们提出了一个多目标优化问题，该问题迫使无人机完成不间断的窃听任务，不进入任何禁飞区（NFS），彼此保持安全距离并尽可能多地收集能量。我们提出了一种基于快速探索随机树（RRT）的路径规划方法，由初步计划阶段和在线计划阶段组成。在在线计划阶段，每个无人机都基于整个团队之间的共享信息来计划自己的路径。由于在计划阶段可以考虑无人机模型，因此无需任何后续调整即可获得所获得的无人机路径。更重要的是，这种方法易于实时实施。