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

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. © 2024 Northeast University. All rights reserved.