A robust controller for non-linear MIMO systems using radial basis function neural network

Abstract

This paper presents a controller design method for multi-input multi-output nonlinear systems using the artificial neural network. The designed controller uses radial basis function networks to generate optimal control signals abiding by constraints, if any, on the control signal, or on the system output. The salient features of the proposed controller include: no requirement of the explicit knowledge of the states of the system, no requirement of a priori knowledge of the non-linear model of the system, and ability of the controller to control time-varying systems or systems having non-minimum-phase characteristics. A recursive algorithm is developed to update the weights of the neural network. Non-linearities in the system as well as variations in system parameters are compensated by the neural network. Simulation results for non-linear systems are included at the end, and controller performance is analysed. A practical case study of single area non-linear automatic generation control is also considered to test the performance of the controller.

Adaptive neural network tracking control of an omnidirectional mobile robot

“…Some works are devoted to the problem of adaptive NN control for all kinds of uncertain nonlinear systems. [31][32][33][34][35] A trajectory tracking controller based on adaptive NN 36,37 is designed for a class of nonlinear discrete-time systems considering dead-zone input. In the work of Boukens and Boukabou, 38 a robust controller based on optimal control theory is presented to solve the trajectory tracking problem of a wheeled mobile robot.…”

Neural network gain-scheduling sliding mode control for three-dimensional trajectory tracking of robotic airships

“…NN can thus serve as black-box models of nonlinear, multivariable static and dynamic systems and can be trained by using input–output data observed on the system. 24 The radial basis function neural network (RBFNN) is one of the most popular NNs and has the advantages of faster learning speed and less chance falling into local minimal convergence. 25 The control gains of NNSSMC can be scheduled synchronously via RBFNN with switching sliding surface and its differential as inputs and control gains as outputs, which can attenuate chattering effectively.…”