Neural network control of flexible-link manipulators using sliding mode

This paper focuses on tracking control problem of flexible-link manipulators. In order to alleviate the effects of nonlinearities and uncertainties, a combined control strategy based on neural network (NN) and the concept of sliding mode control (SMC) is proposed systematically. The chattering phenomenon in conventional SMC is eliminated by incorporated a saturation function in the proposed controller, and the computation burden caused by model dynamics is reduced by applying a two-layer NN with an analytical approximated upper bound, which is used to implement a certain functional estimate. In addition, the Lyapunov analysis can guarantee the signals of closed-loop system bounded and the online NN adaptive laws can make the system states converge to the sliding surface. Furthermore, the boundary layer thickness as well as the gain of corrective control term is also discussed in detail. At last, the theoretic results are validated on the flexible-link manipulator experimental system in Tsinghua University.