柔性空间机械臂RBF神经网络补偿滑模控制策略

RBF Neural Network Compensation Sliding Mode Control Strategy for Flexible Space Manipulators

柔性结构导致柔性空间机械臂的动态参数随着时间产生变化,从而降低了跟踪控制的准确性.质量轻和长径比大导致柔性空间机械臂在运动过程中出现振动现象.为了解决上述问题,本文建立了考虑二维变形和扰动转矩的柔性空间机械臂的动力学模型,推导出简化的非线性动力学方程.在此基础上,设计了控制律,利用RBF(radial basis function)神经网络对柔性空间机械臂中的时变项和扰动转矩进行识别和补偿.然后以双曲正切函数作为逼近率,提出了滑模控制策略.最后,通过仿真和地面物理样机控制实验可以得到,在柔性空间机械臂控制律的设计中,神经网络补偿的控制策略有效地减少了扰动转矩对柔性空间机械臂的影响.并且通过使用tanh函数来代替sgn函数,能够减少输入转矩的波动,更加验证了RBF神经网络补偿滑模控制策略的有效性.

Flexible structures cause the dynamic parameters of flexible space manipulators to change with time,which reduces the accuracy of tracking control.The lighter mass and the larger ratio of length to radius may result in the vibration of flexible space manipulators during their movement.To solve the above problems,a dynamic model of a flexible space manipulator considering two‑dimensional deformation and disturbance torque is established,and a simplified non‑linear dynamic formula is derived.On this basis,a control law is designed to identify and compensate for the time‑varying term and disturbance torque in the flexible space manipulator using the radial basis function(RBF)neural network.Then,using the hyperbolic tangent function as the approximation rate,a sliding mode control strategy is proposed.Finally,through simulation and ground physical prototype experiment,it can be concluded that in the design of control laws for flexible space manipulators,the control strategy with neural network compensation effectively reduces the impact of disturbance torque on the flexible space manipulator.By using the tanh function instead of the sgn function,the fluctuation of input torque can be reduced,and the effectiveness of the RBF neural network compensation sliding mode control strategy is verified.