六足仿生机器人腿部动力学分析与控制  

Dynamics analysis and control of the legs of hexapod biomimetic robot

作　　者：于常娟[1,2] 张明路[1] 金玉阳[1] 

机构地区：[1]河北工业大学机械工程学院,天津300130 [2]廊坊师范学院数学与信息科学学院,河北廊坊065000

出　　处：《河北工业大学学报》2015年第5期28-31,50,共5页Journal of Hebei University of Technology

基　　金：国家自然科学基金(61473113);国家863计划项目(2011AA040201)

摘　　要：为了实现六足仿生机器人的多种运动模式,提高不同路况环境的适应能力,提出了一种带有变形关节的新型腿部结构.采用拉格朗日方法建立了用于控制机器人腿部的动力学模型,设计了计算力矩+RBF神经网络复合控制器,其中计算力矩是已建模的部分,神经网络补偿动力学模型中未建模部分、结构参数的测量误差以及外部扰动.采用MATLAB对腿部的轨迹跟踪控制进行仿真,结果表明了该控制器跟踪精度更高,具有优良控制性能.A new leg structure with a deformable join was presented, so that the hexapod biomimetic robot can achieve multiple motion modes and improve adaptability to environment. By langrange method the dynamics model of the robot legs was obtained. Based on the computed torque and the RBF neural network, the hybrid controller was designed. The computed torque was made of the modeled part. The RBF neural network was used to compensate the unmodeled part, the measure error of the structure parameters and the external disturbance. Simulation results by MATLAB show that the tracking precision of the controller is higher and its control performance is superior.

关 键 词：仿生机器人 新型腿部结构 动力学 计算力矩控制 神经网络 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]