关键词: 机器人阻抗控制, 阻抗模型参数, 神经网络, 机器人力控制

Abstract:

The robot impedance control performance decreases with the efffects of changing or unknown environment stiffness and damping parameters in practical applications,this paper presented a self-adaptive robot impedance control method to resolve this problem. Based on the definition of the robot impedance control performance index and the geometric representation for the impedance model stiffness variation, the calculation method was illustrated for intitial values of the the impedance model stiffness. Designing an artificial neural network to estimate the environmental equivalent stiffness online and combining with critical damping condition of the second-order system, a calculation method was provided for the impedance model damping initial values. The results of robot force control experiments demonstrate  smoother reference trajectory, improved robot force control stability and feasibility in practices compared to the existing robot impedance control methods.

Key words: robot impedance control, impedance model parameter, neural network, robot force contol