无人机多阶段航迹预测协同任务规划

针对多无人机(Unmanned Aerial Vehicles, UAVs)协同控制问题,提出了一种UAVs多阶段航迹预测分布式任务规划方法。定义从一次任务分配开始到其中一项任务完成为一个任务周期。在每个规划周期,首先,各UAV使用A*算法快速预测到所有任务目标的路径,提供至任务分配；然后,采用聚类算法修改目标价值向量,协商分配结果,并实时计算探测范围内的最短路径；最后,采用三次B样条曲线平滑所分配的最短路径,在线规划出满足飞行约束的飞行航迹。通过仿真实验对算法的有效性进行了验证,结果表明,提出的算法能够实时获得近似最优的任务分配结果并规划出可飞行航迹,并有效处理突发任务。

Multi-stage Path Prediction Mission Planning Algorithm for Multiple Unmanned Aerial Vehicles

In this paper, a multi-stage path prediction algorithm of the decentralized mission planning for cooperative UAVs is presented. The planning horizon is defined as the period between the start of task assignment and completion of any task. In every planning horizon, each UAV utilizes the A* algorithm to predict the paths to all tasks and provide the path distances for task assignment. Furthermore, the cluster algorithm is introduced to modify the tasks value vector. The UAVs negotiate the task assignment solution and calculate the shortest path to assigned task in the detection range in real time. Finally, the B-spline curve is addressed to convert the shortest path into flyable smoothing trajectory that subject to the flight constraints. For validation, the scenario of multiple UAVs to perform cooperative missions is considered. Numerical results show that the proposed algorithm can achieve the quasi-optimal assignment solution and generate the flyable trajectory in real time. In addition, the satisfactory performance to accomplish the pop-up tasks is demonstrated.