摘要
针对基于运动相似性的仿人机器人动作规划中固定运动捕获轨迹难以适应可变目标环境的问题,以仿人机器人上阶梯为例,提出了一种基于运动相似性的行走约束与优化控制方法.首先,为相似性阶梯运动轨迹划分子相关键姿势,施加运动学约束;其次,对运动的稳定性、方向、落脚缓冲等施加动力学约束;再次,结合粒子群优化与分层强化学习方法,提出粒子群分层强化矢量位置择优算法,用于对机器人阶梯运动轨迹进行优化.实验证明了方法的有效性.
A method of constraint and optimization control on the similar stepping-upstairs process for humanoid robots is proposed to solve the problem that the fixed tracks captured from the human locomotion can＇t be used in the changeable target environment in the similar motion planning of humanoid robots. Firstly, the similar stepping-upstairs processes are divided into the sub phases and key poses, and the corresponding kinematics constraints are set as well. Secondly, the dynamics constraints are also set for the locomotion stability, walking direction and landing-buffer process. Thirdly, the HRPSO （particle swarm optimization with hierarchical reinforcement learning） algorithm is proposed to optimize the tracks of stepping upstairs for humanoid robots. The experiment proves the validity of the proposed method.
出处
《机器人》
EI
CSCD
北大核心
2014年第2期233-240,共8页
Robot
基金
国家自然科学基金资助项目(61075076
61272382)
广东省自然科学基金资助项目(S2012010009963)
广东省教育厅科技创新项目(2012KJCX0077)
广东高校石化装备故障诊断与信息化控制工程中心项目(512009)
广东石油化工学院博士科研项目启动基金
关键词
仿人机器人
相似性
约束
优化
humanoid robot
similarity
constraint
optimization
