人类控制策略在双足机器人步态控制中的应用  

Application of human control strategy in the gait control of biped robot

作　　者：谭建豪[1] 章兢[1] 何志[1] 

机构地区：[1]湖南大学电气与信息工程学院,长沙410082

出　　处：《电子测量与仪器学报》2013年第1期8-14,共7页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(60634020)资助项目

摘　　要：针对双足机器人面临的复杂环境下动态行走的适应性难题,提出一种基于学习人类控制策略的双足机器人步态控制方法。利用三维线性倒立摆模型构造双足行走系统的状态方程,建立人类控制策略的参数化模型,采用支持向量机(SVM)对人类控制策略进行辨识,并以此SVM作为COM控制器,保证躯干始终处于与地面近似垂直的姿态以及连续的质心运动,增强步态控制的鲁棒性,提高双足机器人在复杂环境下行走的动态稳定性。实验验证了此方法的有效性。Aiming at the adaptability problem about the biped robot walk dynamically in the complex environments, a method of gait control based on learning human control strategy（HCS） is proposed. By means of the status equation of the biped walking system based on the model of 3D linear inverted pendulum, the parametric model of human control strategy is constructed. SVM is used as COM controller, which is adopted to identify human control strategy. The proposed method guarantees that torso is always perpendicular to the ground and the move of center of mass is always continuous. Robust- ness of the gait control is enhanced. The dynamical stability of biped walking in complex environments is improved. The experimental results demonstrate the effectiveness of the proposed methods.

关 键 词：人类控制策略 SVM辨识 双足机器人 步态控制 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TN9[自动化与计算机技术—控制科学与工程]