多目标优化算法在冗余机械手逆解中应用.pdfVIP

第 38 卷增刊 1 中南大学学报( 自然科学版) Vol.38 Suppl.1 2007 年 8 月 J. Cent. South Univ. (Science and Technology) Aug. 2007 多目标优化算法在冗余机械手逆解中的应用 李 胜，申晓宁，陈庆伟，胡维礼 (南京理工大学 自动化学院，江苏 南京，210094) 摘 要：针对带有 N 个关节的冗余机械手运动学逆解问题，提出了一种改进的多目标优化遗传算法。与传统的机 械手逆解方法相比，所提算法不仅考虑了使终端执行器精确地到达期望位置的目标，而且同时优化了关节转动角、 机械手柔顺性、与障碍物之间的距离 3 个目标。在算法群体的初始化及子代群体的生成中，采用了一种新型的个 体生成方式以满足约束条件的限制；在选择算子中，引入了防止算法早熟收敛的机制。仿真结果表明，所提算法 能够有效地求解具有多个优化目标的机械手运动学逆解问题。 关键词：多目标优化；遗传算法；冗余机械手；运动学逆解 中图分类号：TP24 文献标识码：A 文章编号：1672-7207(2007)S1−0598−05 Application of multi-objective optimal algorithm in kinematical inverse of hyper redundant manipulator LI Sheng, SHEN Xiao-ning, CHEN Qing-wei, HU Wei-li (School of Automation, Nanjing University of Science and Technology, Nanjing 210094, China) Abstract: An improved multi-objective optimization genetic algorithm was proposed for solving kinematical inverse solutions of the redundant manipulator. Compared with the traditional methods, the proposed algorithm not only made the end-effector reach the desired position precisely, but also optimized three objectives that were turning angles of joints, flexibility of the manipulator and the distance to obstacles. A new method to create individuals satisfying the constraints was proposed in the initialization of the algorithm and in generating child population, and a mechanism that prevented premature of the algorithm was incorporated in the selection operator. Simulation results indicate that the proposed algorithm can solve kinematical inverse solutions of the manipulator with multiple optimization objectives effectively. Key words: multi-o