考虑摩擦特性的柔性关节空间机器人自适应CMAC神经网络鲁棒控制

Adaptive CMAC neural network robust control for flexible joints space robot with friction characteristics

讨论了关节摩擦力矩影响下,具有柔性铰关节的漂浮基空间机器人系统的动力学控制问题。设计了基于高斯基函数的小脑神经网络(CMAC)鲁棒控制器和摩擦力矩补偿器。用奇异摄动理论对系统的动力学模型进行快慢变子系统分解,针对快变子系统,设计力矩微分反馈控制器来抑制机械臂关节柔性引起的振动;对于慢变子系统,设计了基于自适应CMAC神经网络鲁棒控制器以实现系统参数不确定情况下的关节轨迹跟踪,并设计基于摩擦力上界的补偿器消除摩擦力矩影响。与传统的CMAC神经网络控制相比,该控制器能有效改善非线性关节摩擦引起的迟滞问题,具有快速准确跟踪期望轨迹的能力。运用Lyapunov稳定性理论证明了该控制器的稳定性,仿真结果证明该方法有效。

The dynamic control of a floating space robot system with flexible joints under the influence of joint friction torques is discussed.A CMAC robust controller based on Gaussian basis function and friction torque compensator are proposed.The singular perturbation theory is used to decompose the dynamic model of the system into fast and slow subsystems.A torque differential feedback controller is designed for the fast subsystem to suppress the vibration caused by the joint flexibility of the manipulator.For the slow subsystem, a robust controller based on adaptive CMAC neural network is designed to realize the joint trajectory tracking under the condition of uncertain parameters of the system, and a compensator based on the upper bound of friction force is designed to eliminate the influence of friction torque.Compared with the traditional CMAC neural network control, the proposed controller can effectively improve the hysteresis problem caused by nonlinear joint friction, and has the ability to track the desired trajectory quickly and accurately.The stability of the controller is proved by Lyapunov stability theory.The effectiveness of the scheme is verified by the simulation results.