复杂环境下群机器人自组织协同多目标围捕

针对动态多目标围捕,提出了一种复杂环境下协同自组织多目标围捕方法.首先设计了多目标在复杂环境下的运动模型,然后通过对生物群体围捕行为的研究,构建了多目标简化虚拟受力模型.基于此受力模型和提出的动态多目标自组织任务分配算法,提出了群机器人协同自组织动态多目标围捕算法,这两个算法只需多目标和个体两最近邻位置信息以及个体面向多目标中心方向的两最近邻任务信息,计算简单高效,易于实现.接着获得了系统稳定时参数的设置范围.由仿真可知,所提的方法具有较好的灵活性、可扩展性和鲁棒性.最后给出了所提方法相较于其它方法的优势.

Self-organizing cooperative multitarget hunting by swarm robots in unknown dynamic complex environments

A self-organizing method based on a multitarget simplified virtual-force model is proposed for nonholonomic mobile swarm robots hunting in unknown dynamic environments. First, the motion equations of individual robots are given; then the motion models for the hunting multitarget and dynamic obstacles in unknown dynamic environments are presented. Through the decomposition of hunting behavior under complicated environments, a multitarget simplified virtual-force model is formed. Based on the virtual-force model, a dynamic multitarget self-organizing task assignment algorithm and a cooperative self-organizing dynamic multitarget hunting algorithm are designed. The two algorithms only need the position information of multitarget and the two nearest neighbors, and the task information of two nearest neighbors to the direction for the multitarget center. The new method is simple and efficient, and easy to implement. After that, the stability of the hunting system is analyzed and the control parameter ranges are given. Simulation results demonstrate that the proposed self-organizing cooperative multitarget hunting method can make the group of robots maintain a good hunting formation for hunting multitarget in unknown dynamic obstacles environments and has good performance of obstacles avoidance, robustness, scalability, and flexibility. Finally, some advantages of this hunting method are presented, compared with the hunting method based on other methods.