Abstract: Aiming at the optimization problems of complex multi-robot task allocation,there are always high-dimensional motion planning calculations and insufficient applicability of planning results in a dynamic environment.Taking the multi-robot inspection station for an example,a hierarchical task allocation method was proposed for multi-robot coordinated motion planning to reduce the possibility of interference in the multi-robot shared space.For the shared task set of multiple robots,a task allocation method based on the lazy traveling salesman problem solution was proposed to realize the allocation of a large number of task sets to multiple robots.Then fuzzy C-Means was used to cluster the single robot task set.On the basis of clustering,a multi-objective task allocation optimization model considering the center of the subset clustering and the motion time of the single robot was proposed to realize the task allocation to multiple robots in the sensing process of multiple products.Through an inline inspection case of an auto body,the task allocation method was verified in terms of inspection cycle and multi-robot consistency.Results showed that the proposed method could realize efficient allocation of tasks for inline measurement,and could reduce the conflict probability of multi-robot effectively.

Key words: auto body, quality inspection, multi-robotics, task allocation, path planning

摘要： 针对动态环境下多机器人任务分配过程高维运动规划计算量大、优化分析复杂与结果适用性不足等问题,以多机器人检测工位为对象,提出面向多机器人协调运动规划的层级化任务分配方法,降低任务分配计算量、减少共享空间内多机干涉可能性。首先,针对多机器人共享任务集,提出一种基于惰性旅行商问题求解的粗分配方法,实现大量任务集到多机器人的分配。然后,采用模糊C-Means对单机器人任务集进行聚类;在此基础上,建立了面向任务子集聚类中心和单机运动时间等多目标的任务优化模型,实现测量任务到多车次、多机的细化分配。最后,通过某车型在线测量案例,从测量周期、多机一致性等方面对所提分配方法进行验证。结果表明,提出的层次化任务分配方法可实现在线测量下多任务的高效分配,同时能有效降低多机器人运动干涉概率。

关键词: 车身, 质量检测, 多机器人, 任务分配, 路径规划