Vibration Suppression of Flexible Joints Space Robot based on Neural Network

A vibration suppression control method based on neural network for space robot is proposed. At the same time, considering the uncertain factors such as model inaccuracy and joint flexibility, the dynamic model of space robot is established. Based on singular perturbation theory, it is decomposed into fast subsystem and slow subsystem models for control respectively. Using the nonlinear approximation characteristics of neural network, a radial basis function neural network (RBFNN) controller is designed to approximate the unknown nonlinear system. In order to obtain the best hidden layer Gaussian map, the Taylor linearization method is used to linearize the nonlinear Gaussian function of RBFNN, and the on-line adaptive adjustment rates of parameters including network weight, Gaussian function center and width are designed to improve the approximation accuracy of neural network. A robust controller is designed to suppress the external disturbances. A velocity differential feedback controller (VDFC) is designed to suppress the elastic vibration of flexible joints. Experimental results show the effectiveness of the proposed scheme. © 2022, IAENG International Journal of Applied Mathematics. All Rights Reserved.