Abstract: The task allocation and process planning for the multi-station &multi-robots in the spot welding process are key issues for the assembly efficiency.This problem involves complicated sub-problems such as multi-station and multi-robot hierarchical task allocation,single-robot spot welding sequence planning,and multi-robot collaboration in single station.It is a highly coupled optimization problem with complex engineering constraints.Traditional task allocation methods are often solved by step-by-step optimization,and it is difficult to obtain optimal task planning results in a multi-station multi-robot system.To address this problem,considering multiple constraints such as robot reachability,collision detection,welding cycle time within the station and multi-robot consistency,a comprehensive optimization model was constructed for the Multi-Station Multi-Robot Task Allocation (MS-MRTA) problem.Furthermore,an improved self-organizing neural network and a cyclic optimization strategy were proposed for optimal task allocation of MS-MRTA problem.To evaluate the effectiveness of the proposed method,a body side assembly case was used for comparative analysis with different state-of-art task allocation methods.Results showed that the multi-station assembly cycle was reduced by 14.49% on average based on the proposed algorithm,and the consistency of multiple robots in the stations was also improved by 54.22% on average.

Key words: auto body, spot welding, task allocation, multi-robot, intelligent optimization

摘要： 整车焊装过程中多工位多机器人的任务分配与工艺规划是影响装配效率的关键问题。该问题涉及到多工位多机器人的层级化任务分配、单机器人点焊次序规划以及多机器人协同等子问题,是一个高度耦合且具有复杂工程约束的优化问题。传统任务分配方法往往以分步优化方式进行求解,在多工位多机器人系统中难以获得有效任务规划结果。为了解决该问题,考虑了机器人可达性、碰撞检测、工位内焊接周期等多重约束,构建了面向层级化任务分配与焊接次序协同规划的多工位-多机器人任务分配(MS-MRTA)问题的综合优化模型,提出改进的自组织神经网络及循环优化策略,获得了MS-MRTA问题的优化求解方案。通过整车侧围案例对所提出方法进行应用验证,结果表明,所提出算法将多工位焊装周期平均下降14.49%,工位内多机器人运行时间一致性提升54.22%。

关键词: 车身, 焊装工艺, 任务分配, 多机器人, 智能优化