采摘机械手的路径规划——基于相似矢量矩免疫遗传算法

Path Planning of Picking Manipulator——Based on Similarity Vector Moment and Immune Genetic Algorithm

为了提高果蔬采摘机器人机械手采摘路径规划决策的效率,降低数据分析的冗余性,在分析生物免疫系统功能、作用机理及特点的基础上,提出了一种基于相似性矢量距的免疫遗传算法,并将其植入到了机械手躲避障碍物的决策系统中。该算法利用相似矢量矩方法将选择概率的局部搜索和整体搜索有机地结合起来,使算法保留了最优抗体,保证了计算不陷入局部最优解,并使抗体的计算具有了多样性。为了验证算法的有效性和可靠性,设计了果实采摘机器人的测试样机。通过测试发现:机器人机械手动作过程中可以有效地躲避障碍物,到达指定的目标采摘区域,在相似矢量矩免疫遗传算法、免疫遗传算法和相似矢量矩3种算法的测试结果中,相似矢量矩免疫遗传算法的收敛性最好,优化路径最优。

In order to improve the efficiency and reduce the redundancy of data analysis,a new immune genetic algorithm is proposed based on similarity vector distance and the analysis of the function and mechanism of biological immune system. In this algorithm,the local search and global search of the selection probability is combined with the method of similarity vector moment,which makes the algorithm retain the optimal antibody,and ensure that the computation does not fall into local optimal solution. In order to verify the validity and reliability of the algorithm,the test prototype of the fruit picking robot is designed. The algorithm can effectively avoid obstacles and reach the designated target picking area. In the test results of three kinds of algorithm,the immune genetic algorithm,immune genetic algorithm and similarity vector moment algorithm,the optimal path is optimized.