摘要
针对仿生控制的特点,构建了一种Matlab环境下的双足机器人动力学仿真及仿生控制平台。为了充分利用关节自身的被动特性,描述了由非外界约束引起的自由度变化,以及双足机器人在运动相态间的合理切换。同时,考虑到碰撞是机器人运动的固有特性,建立了包括碰撞在内的混合动力学模型,最终实现控制器与机器人、环境的耦合。在此平台上,对设计的仿生控制算法进行了仿真:借助机器人自身运动状态和环境交互信息,协调各个关节的动作时序,实现了低能耗的稳定行走。结果表明,该平台为双足机器人仿生控制及其运动特性分析提供了有效的仿真手段。
Aiming at bionic control,a dynamics simulation and bionic control platform for biped robot was established based on Matlab.To utilize passive locomotive characteristics of joints,the variation of degrees of freedom caused by non-environmental constraints,and the consequent transformations between different motion phases were studied.Considering impact as inevitable character of robot locomotion,the hybrid dynamics model including impact was established,and the entrainment between controller,robot,and environment were finally realized.A bionic control method was proposed and tested on this platform.Depending on states of robot and environment information,the locomotion of each joint was coordinated,and the robot could walk stably with low energy consumption.The simulation result shows this platform provides an effective simulation means for bionic control and its motion characteristics analysis.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2011年第5期977-983,共7页
Journal of System Simulation
基金
国家自然科学基金(50575119
50805082)
国家863计划(2006AA04Z253)
清华大学摩擦学国家重点实验室自主研究课题(SKLT08B09)
作者简介
王健美（1981-），女，四川人，博士生，研究方向为机器人与仿生学；
陈恳（1954-），男，四川人，博士，教授，研究方向为机器人与仿生学，特种机器人应用，机电自动化及系统集成。
