Abstract: The assembly interfaces are critical machining features ensuring the assembly quality for large cylindrical components.It is necessary to use effective measurement methods to obtain machining allowance and detect machining quality.The traditional measurement methods are characterized by complex processes,a heavy reliance on manual experience,and low efficiency.To address this problem,a robot-assisted laser scanning method for the assembly interfaces of large cylindrical components was proposed as an example of a large cylinder measurement process,and a measurement path planning study was carried out.Based on clustering algorithms,a viewpoint planning method for scanning measurement of assembly interfaces was proposed.Additionally,a viewpoint adjustment method for determining and resolving spatial interference and occlusions in the field of view was introduced.127 interference-free scanning measurement viewpoints had been planned.The Pose and Obstacle-aware Variable Weight Ant Colony Optimization (POVW-ACO) algorithm was proposed for path planning of scanning measurement viewpoints.The result and robustness were better than the traditional ACO algorithm.The planned scanning path was verified by simulation and physical experiment,and the results met the expected requirements in about 4.5 min.

Key words: large-scale component, robotic scanning measurement, path planning, ant colony algorithm

摘要： 对于大型构件而言,装配界面是保证产品最终装配质量的关键加工特征,其测量加工一体化是一类典型且重要的数字化与智能制造场景,通过有效的方法获取加工余量并检测加工质量尤为重要。传统测量模式工艺复杂,人工经验依赖性强,测量效率低。针对这一问题,以某大型筒件测量工艺为例,提出一种面向大型筒件装配界面的机器人辅助激光扫描方法,并开展测量路径规划研究。首先,基于聚类算法提出面向装配界面的扫描测量视点规划方法,并提出判断及处理空间干涉和视场被遮挡的视点调整方法,规划出127个无干涉扫描测量视点。其次,提出考虑姿态变化及障碍物的可变权重因子蚁群算法对扫描测量视点进行路径规划,求解结果和鲁棒性均优于传统蚁群算法。最后,对规划的扫描测量路径进行仿真验证及实物实验验证,在约4.5 min的时间内获取到满足预期要求的扫描测量结果。

关键词: 大型构件, 机器人扫描测量, 路径规划, 蚁群算法