Gottlieb Yoav, Shima Tal
Technion-Israel Institute of Technology, Technion City, Haifa 3200003, Israel.
Sensors (Basel). 2015 Nov 24;15(11):29734-64. doi: 10.3390/s151129734.
The intertwined task assignment and motion planning problem of assigning a team of fixed-winged unmanned aerial vehicles to a set of prioritized targets in an environment with obstacles is addressed. It is assumed that the targets' locations and initial priorities are determined using a network of unattended ground sensors used to detect potential threats at restricted zones. The targets are characterized by a time-varying level of importance, and timing constraints must be fulfilled before a vehicle is allowed to visit a specific target. It is assumed that the vehicles are carrying body-fixed sensors and, thus, are required to approach a designated target while flying straight and level. The fixed-winged aerial vehicles are modeled as Dubins vehicles, i.e., having a constant speed and a minimum turning radius constraint. The investigated integrated problem of task assignment and motion planning is posed in the form of a decision tree, and two search algorithms are proposed: an exhaustive algorithm that improves over run time and provides the minimum cost solution, encoded in the tree, and a greedy algorithm that provides a quick feasible solution. To satisfy the target's visitation timing constraint, a path elongation motion planning algorithm amidst obstacles is provided. Using simulations, the performance of the algorithms is compared, evaluated and exemplified.
解决了在存在障碍物的环境中，为一组固定翼无人机分配一组优先目标的任务分配与运动规划相互交织的问题。假设目标的位置和初始优先级是通过用于检测受限区域潜在威胁的无人地面传感器网络确定的。目标具有随时间变化的重要性级别，并且在车辆被允许访问特定目标之前必须满足时间约束。假设车辆携带机身固定传感器，因此需要在直线平飞的同时接近指定目标。固定翼飞行器被建模为杜宾斯飞行器，即具有恒定速度和最小转弯半径约束。所研究的任务分配与运动规划的综合问题以决策树的形式提出，并提出了两种搜索算法：一种详尽算法，它在运行时间上有所改进并提供树中编码的最小成本解决方案；以及一种贪婪算法，它提供快速可行的解决方案。为了满足目标的访问时间约束，提供了一种在障碍物间的路径延长运动规划算法。通过仿真，对算法的性能进行了比较、评估和举例说明。
