• CN:11-2187/TH
  • ISSN:0577-6686

›› 2006, Vol. 42 ›› Issue (增刊): 35-38.

• 论文 • 上一篇    下一篇

扫码分享

基于遗传算法的三肢体功能仿生机器人步态能量优化

庞明;赵杰;蔡鹤皋   

  1. 哈尔滨工业大学机器人研究所
  • 发布日期:2006-12-30

GAIT ENERGY OPTIMIZATION OF THREE-LIMBED ROBOT BASED ON GENERATION ALGORITHM

PANG Ming;ZHAO Jie;CAI Hegao   

  1. Robotics Institute, Harbin Institute of Technology
  • Published:2006-12-30

摘要: 针对实验室研究设计的三肢体功能仿生机器人所要实现的基本行走功能而提出五种基本行走步态:平行交叉行走步态、翻转行走步态、高姿爬行行走步态、低姿爬行行走步态和两肢体行走步态。在分析了机器人运动学和动力学模型的基础上,给出了生成机器人基本行走步态各关节轨迹的计算方法。随后针对这五种基本行走步态提出一种以步行速度和能量消耗为目标函数的步态优化方法。该方法采用遗传算法进行优化,并通过最佳保留、对操作结果的取舍和动态调整变异概率三项措施有效地避免了其早熟的问题,得到了较为可靠的结果。最后提出嵌套优化的策略,在消耗时间相同的前提下,提高了遗传算法的计算精度。两组随机试验结果的对比表明了该嵌套优化策略有效地提高了优化结果的质量。

关键词: 步态优化, 功能仿生, 嵌套优化, 三肢体

Abstract: Five kinds of basic gaits of a three limbed robot are introduced and a gait optimization method of these basic gaits is discussed in detail. By the kinematical and dynamitic analysis, an objective function of gait optimization is set up which can be used to optimize the velocity and power of the three-limbed robot. Then an improved program of generation algorithm is studied which can avoid the problem of premature convergence and obtain credible results. A method of nested optimization of generation algorithm is also developed which improves the precision of generation algorithm with consuming the same time and then the random experiment results performed have proved that the method is effective.

Key words: Gait optimization, Functional bionics, Nested optimization, Three-limbed

中图分类号: