Auto-tuning control techniques applied to a neonatal incubator

Abstract

This work studies and proposes techniques of automatic tuning of controllers for single-inputsingle-output (SISO) and two-inputs-two-outputs (TITO) processes with dead time. Furthermore, it proposes a dead-time compensator (DTC) based on generalized predictive control (GPC) with anti-windup action. Firstly, generalities about the studied auto-tuning methodology are presented. The auto-tuning of proportional-integral-derivative (PID) controllers for SISO processes is studied e new methods for integrating and unstable processes are proposed. Following, it is proposed the auto-tuning of the simplified filtered Smith predictor, a robust dead-time compensator for stable, unstable and integrating SISO processes. The results for SISO processes are then extended for the auto-tuning of PID controllers for TITO processes. Finally, the GPC-based DTC with anti-windup action is presented. Its formulation can cope with stable, unstable and integrating SISO processes. Simulation results are presented, where studied and proposed methods are compared with other recent methods of the literature, highlighting their properties and advantages. Experimental results for stable SISO and TITO processes were obtained by applying the studied methods on the control of temperature and relative humidity of a neonatal incubator prototype and verify the effectiveness of the studied and proposed methods in a practical context.