Sliding mode control of parallel robot by optimizing switching gain based on RBF neural network

The parallel robot system, consisted of parallel motion mechanism with multiple linkages, has the advantages of high rigidity, strong bearing capacity and so on. To the control problem of parallel robot with multi-variables, non-linearity and strong coupling, a novel sliding mode control method which uses RBF neural network to optimize its switching gain is proposed for a 2-DOF redundant parallel robot. Firstly, the parallel robot kinematics analysis is completed by the geometric relationship of the parallel mechanism; Secondly, the mathematical model of the control branch is established based on the decoupling characteristics of sliding mode control design; then, by adjusting the weights of RBF neural network based on the gradient descent, the design of sliding mode control algorithm which uses RBF neural network to optimal switching gain is realized and its stability is theoretically proved. The simulation and experimental results show that the control method dose not need accurate mathematical mode of controlled object and is not sensitive to uncertain parameter and external disturbance variations. It is shown from the comparison of the simulation and experimental results with the fixed-gain sliding mode control that the proposed RBF neural network sliding mode control has smoother control volume, and can achieve the higher performance control of parallel robot system.