论文部分内容阅读
为实现机器人关节位置镇定和轨迹跟踪控制,控制律的设计须针对确定的机器人动力学模型,由于机器人结构参数、作业环境的外界干扰及结构振动等不确定性因素的存在,会造成机器人动力学模型不确定.为此,设计3个RBF神经网络分别对不确定机器人模型中的3个不确定项进行分块建模,得到机器人估计模型,神经网络的权值采用自适应算法.针对机器人估计模型设计PI鲁棒滑模控制律.将所设计的控制器用于三关节机器人的三个关节的力矩控制,研究结果表明:三关节均约在1 s时达到期望位置和跟踪期望轨迹,镇定误差和跟踪误差也快速、稳定地趋于零.通过定义基于积分型的Lyapunov函数,利用Lyapunov稳定性理论证明了控制系统是全局渐近稳定的.
In order to realize the stabilization of the joint position of the robot and the trajectory tracking control, the design of the control law must be based on the determined robot dynamics model. Due to the uncertainties such as structural parameters of the robot, external disturbance of the working environment and structural vibration, the robot dynamics The model is indefinite.Therefore, three RBF neural networks are designed respectively to block modeling of three uncertain items in the uncertain robot model, the robot estimation model is obtained, and the weight of the neural network is adaptive.According to the robot estimation The model is designed to be PI robust sliding mode control law. The designed controller is applied to the moment control of three joints of the three-joint robot. The results show that the desired joint position and tracking desired trajectory are achieved at about 1 s for each joint, And the tracking error quickly and stably approaches zero.By defining the Lyapunov function based on integral type, the Lyapunov stability theory is used to prove that the control system is globally asymptotically stable.