基于马尔可夫决策过程的六足机器人自由步态规划
基于马尔可夫决策过程的六足机器人自由步态规划李满宏1，张建华1，张小俊1,2，张明路1（1. 河北工业大学机械工程学院，天津300130； 2. 哈尔滨工业大学机器人技术与系统国家重点实验室，黑龙江哈尔滨150080）摘要：为精细模仿生物步态，充分发挥六足机器人运动潜能，本文在离散化机器人足端轨迹的基础上，融合中枢模式发生器（CPG）模型与反射模型，建立了离散化步态模型，基于稳定性分析，构建了机器人稳定的位置状态空间，将复杂的步态规划问题等效转化为稳定的位置状态空间中位置状态间的排序问题，在此基础上，提出了一种自由步态生成算法；并基于处理顺序决策问题的马尔可夫决策过程，以平均稳定裕量为优化指标，针对特定地形研究自由步态的优化算法．样机步态实验结果表明，自由步态生成算法与优化算法均可生成在一定程度上符合生物运动特点的稳定步态，且自由步态优化算法可针对特定地形快速规划出基于平均稳定裕量的最优步态．
关键词：六足机器人；自由步态；步态规划；马尔可夫决策过程；离散化
中图分类号：TP242 文献标识码：A 文章编号：1002-0446(2015)-05-0529-09
Free Gait Planning for a Hexapod Robot Based on Markov Decision ProcessLI Manhong1 ，ZHANG Jianhua1 ，ZHANG Xiaojun1,2 ，ZHANG Minglu1(1. School of Mechanical Engineering, Hebei University of Technology, Tianjin 300130, China;2. State Key Laboratory of Robotics and System, Harbin Institute of Technology, Harbin 150080, China)
Abstract: In order to imitate biological gait accurately and develop the movement potential of hexapod robots comprehensively,a discrete gait model is built based on the discretization of foot trajectories and the fusion of CPG (central pattern generator) model and reflect model. Firstly, the stable position state space is constructed based on stability analysis, and the complex gait planning is transformed into the sequencing problem of position states in the stable position state space. Then,a free gait generation algorithm is proposed, and the optimized free gait planning algorithm for specific terrain is investigated by taking the average stability margin as performance index based on Markov decision process, which is good at dealing with sequential decision problem. The gait experiment results of the prototype show that both the free gait generation algorithm and the optimized algorithm can generate stable gait which accords with motion characteristics of creatures to some extent,and the optimized free gait planning algorithm can plan the optimized gait for specific terrain based on average stability margin quickly.
Keywords: hexapod robot; free gait; gait planning; Markov decision process; discretization
页:
[1]