摘要
针对仿人机器人机械连杆控制结构的动力学、运动学等特点,建立了25自由度仿人机器人家谱树形结构,提供了连杆数据列表,分析了步态规划约束、步行倒立摆、奇异位姿控制、落地步态方向控制等特性,构建了行走步态模型,并采用链接图法对环境空间建模,引入种群规模自调整遗传算法实现路径规划。以仿人机器人NAO为实验平台,采用链接图法进行全局建模,实验表明,改进的遗传算法在仿人机器人路径规划中具有更快的收敛速度与更好的适应性。
Aimed at the characteristics of control structure, dynamics and kinematics of the mechanical links in hu- manoid robot,authors set up the genealogy tree structure with 25 degrees of freedom,and the data list of links,ana- lyzed such characteristics as the step planning constraint, walking inverted pendulum,odd gesture control and landing direction of step control and constructed the step walking model. The linking graph was used to construct the envi- ronment space, and the genetic algorithm with self-adjusting group size was used to realize the path planning. The ex- periment proved that the improved genetic algorithm converged faster and was more adaptive than the traditional one by taking the humanoid robot NAO as the platform and using the linking graph to carry out the global modeling.
出处
《山东科技大学学报（自然科学版）》
CAS
2012年第2期86-92,共7页
Journal of Shandong University of Science and Technology(Natural Science)
基金
国家高技术研究发展计划("863"计划)项目(2007AA041603)
国家自然科学基金项目(60905047
60643005)
广东高校石化过程装备故障诊断与信息化控制工程中心重点实验室项目
广东省自然科学基金项目(10252500002000001)
关键词
仿人机器人
约束
零力矩点
路径规划
humanoid robot
constraint
zero moment point
path planning
