Error model-oriented vibration suppression control of free-floating space robot with flexible joints based on adaptive neural network

Abstract

At the same time, considering the uncertain factors such as load variation, external interference and joint flexibility in engineering practice, a robust control method based on adaptive neural network is proposed. The dynamic model of free-floating space robot is established, and the error model caused by uncertain factors is deduced. Different from the traditional compensation algorithm that ignores the error model, a compensation controller based on Radial basis function neural network (RBFNN) is designed to approximate the error model. The approximation error is eliminated by robust controller to improve the control accuracy. In order to make full use of the nonlinear approximation ability of neural network, the error model is decomposed into four parts according to the input characteristics, and the neural network compensator is designed for separate and overall compensation, which further improves the control accuracy and robustness. The adaptive learning rates of network weights are designed to ensure online real-time adjustment without offline learning stage. A flexible compensator based on torque and a controller based on moment difference feedback controller (MDFC) are designed to suppress elastic vibration. Simulation and experimental studies show that the proposed strategy can have good compensation performance and robustness, and can better suppress elastic vibration, which proves the effectiveness and superiority of the proposed scheme.