Adaptive approximation-based design mechanism for non-strict-feedback nonlinear MIMO systems with application to continuous stirred tank reactor

摘要

This article is concerned with the problem of adaptive neural controller design for multi-input/multioutput nonlinear systems with input-saturations and disturbances. In the proposed design mechanism, we will take advantage of hyperbolic tangent functions to smooth the sharp corners of the input saturations and use Young's inequality to handle the nonlinear terms derived from the deducing process, and meanwhile apply the intelligent algorithm to estimate the unknown nonlinearity via neural networks. Furthermore, the backstepping technique is used to complete the design of the controller and Lyapunov stability theory is employed to show that the whole closed-loop system is semi-global uniformly ultimately bounded and the tracking error is bounded subject to the small neighborhood of the origin. Finally, as a practical application of the researched design scheme, adaptive neural controller for a continuous stirred tank reactor is constructed. (C) 2019 ISA. Published by Elsevier Ltd. All rights reserved.