An intelligent sliding mode controller based on LAMDA for a class of SISO uncertain systems

This paper presents a new intelligent sliding mode controller based on LAMDA (Learning Algorithm for Multivariate Data Analysis), a fuzzy method used for supervised and unsupervised learning applicable to the detection of functional systemic states. LAMDA computes the Global Adequacy Degree (GAD) of an object to a class or functional state to determine its degree of similarity. An inference stage has been added to LAMDA to make it work as a controller, in combination with the basic features of a sliding mode control (SMC) and Lyapunov stability theory. The novelty of this proposal is that we have used the LAMDA algorithm to compute the SMC continuous and discontinuous control actions to obtain a chattering-free controller, which can then be applied to a class of SISO systems with variable dynamics and model uncertainties. Simulations on two nonlinear chemical processes have validated the proposal: 1) control of a continuous stirred tank reactor (CSTR) under bounded disturbances and reference changes, and 2) regulation of a mixing tank with variable parameters (variable dynamics). The experiments are compared with other control techniques, demonstrating that the proposed method can accurately control the tanks, improving the results in performance, robustness, and disturbance rejection. (c) 2021 Elsevier Inc. All rights reserved.

Automated summary

This paper presents a new intelligent sliding mode controller based on the LAMDA algorithm for computing continuous and discontinuous control actions, which can be applied to SISO systems. Simulations on nonlinear chemical processes validate the method's excellent performance in controlling tanks, significantly improving performance and robustness.