Weighted Multiple Neural Network Boundary Control for a Flexible Manipulator With Uncertain Parameters

This paper addresses the angle tracking and vibration suppression for a flexible manipulator with uncertain parameters. Based on the partial differential equation (PDE) model, a unified framework of weighted multiple neural network boundary control (WMNNBC) is proposed to deal with the jumping parameters, in which neural networks are designed as the local boundary controllers to suppress vibrations. A novel proportion-derivative-like machine learning algorithm is developed to guarantee the learning convergence. Besides, the weighting algorithm is used to fuse multiple local neural network controllers to generate the appropriate global control signals with the variations of plant parameters. The stability of the overall closed-loop system is proved by the virtual equivalent system (VES) theory. The simulations are implemented to illustrate the feasibility and control performance of the proposed WMNNBC strategy.