Abstract:

A method of sliding mode control based on SVM(Support Vector Machine)nonlinear estimation of the robotic dynamics has been proposed to acquire faster trajectory tracking rate of the space manipulator end. SVM regression is used to estimate the parameter uncertainty of the nonlinear term and to compensate dynamic model of space robot. This SVM nonlinearity compensated sliding mode controller designed on the stability of the closed-loop system, avoiding computing precise model, has improved the dynamics performance of the system, achieving faster target trajectory tracking rate and reducing the tracking error. The simulation results show that the presented method is effective and feasible.

Key words: space robot, support vector machine（SVM）, nonlinear estimation, sliding mode control