论文部分内容阅读
针对传统机器人零件微装配算法不能有效解决干扰区域划分问题,提出一种模糊优化结合机器学习的智能干扰区域划分算法。首先,根据工作区内零件与目标孔之间的位置关系设计模糊协调器和特殊的规则库,避免激活干扰状态。然后,利用提出的智能区域划分算法合并所有相邻区域。最后,通过决策制定,选择区域中最低模糊熵的四元组控制值作为执行分配任务的最终控制值。实验结果表明,该算法可将69个子区域合并降至10个子区域,相比其他的较为先进的装配算法,算法更加灵活,显著提高了任务效率。
Aiming at the problem that the traditional robotic assembly micro-assembly algorithm can not effectively solve the problem of interference area division, this paper proposes a fuzzy optimization algorithm based on machine learning. First, the fuzzy coordinator and the special rules base are designed according to the position relationship between the parts and the target hole in the work area to avoid activating the interference state. Then, all the adjacent regions are merged using the proposed smart region demarcation algorithm. Finally, through decision-making, the quadtree control value of the lowest fuzzy entropy in the region is selected as the final control value for performing the assigned task. Experimental results show that the algorithm can reduce 69 sub-regions to 10 sub-regions. Compared with other more advanced assembly algorithms, the algorithm is more flexible and significantly improves the task efficiency.