华南理工大学学报(自然科学版) ›› 2011, Vol. 39 ›› Issue (10): 68-73.doi: 10.3969/j.issn.1000-565X.2011.10.012

• 计算机科学与技术 • 上一篇    下一篇

基于强度Pareto 进化算法的双足机器人步态规划

毕盛1 庄钟杰1 闵华清2   

  1. 1.华南理工大学 计算机科学与工程学院,广东 广州 510006; 2.华南理工大学 软件学院,广东 广州 510006
  • 收稿日期:2010-12-17 修回日期:2011-07-14 出版日期:2011-10-25 发布日期:2011-09-01
  • 通信作者: 毕盛(1978-) ,男,讲师,博士,主要从事仿人机器人研究. E-mail:picy@ scut.edu.cn
  • 作者简介:毕盛(1978-) ,男,讲师,博士,主要从事仿人机器人研究.
  • 基金资助:

    广东省自然科学基金资助项目( S2011040002784)

Gait Planning of Biped Robots Based on Strength Pareto Evolutionary Algorithm

Bi ShengZhuang Zhong-jieMin Hua-qing2   

  1. 1. School of Computer Science and Engineering,South China University of Technology,Guangzhou 510006,Guangdong,China; 2. School of Software Engineering,South China University of Technology,Guangzhou 510006,Guangdong,China
  • Received:2010-12-17 Revised:2011-07-14 Online:2011-10-25 Published:2011-09-01
  • Contact: 毕盛(1978-) ,男,讲师,博士,主要从事仿人机器人研究. E-mail:picy@ scut.edu.cn
  • About author:毕盛(1978-) ,男,讲师,博士,主要从事仿人机器人研究.
  • Supported by:

    广东省自然科学基金资助项目( S2011040002784)

摘要: 为了获得良好的双足机器人步行模式,提出了以步行过程中机器人的稳定性、移动性和能耗为目标的步态规划多目标优化方法.该方法基于倒立摆模型产生基本步态,并使用罚函数法和改进的强度Pareto 进化算法( SPEA2) 在可行域中求得基于基本步态的Pareto 解集,从而找出最优解.最后在Matlab6.5 仿真环境下进行步态仿真,并将产生的步态应用于SCUT-I 型仿人机器人,实现了平均步行速度为0. 26m/s 的稳定行走.

关键词: 仿人机器人, 步态规划, 多目标进化算法, 强度Pareto 进化算法

Abstract:

In order to achieve a good walking pattern of biped robots,a multi-objective optimal method for the gait planning is proposed,with the stability,the mobility and the energy of the robot as the research focuses. In this method,the basic gaits are generated based on the inverted pendulum. Then,Pareto optimal solutions based on the basic gaits are obtained in the feasible region by means of the improved strength Pareto evolutionary algorithm ( SPEA2) and the penalty function method. Finally,after the walking simulation through Matlab 6.5,some gaits are generated and are then used in SCUT-I Humanoid Robot,thus obtaining a stable walking pattern with an average velocity of 0. 26m/s.

Key words: humanoid robot, gait planning, multi-objective evolutionary algorithm, strength Pareto evolutionary algorithm

中图分类号: