New automatic tuning of multivariable PID controller applied to a neonatal incubator

This work studies an automatic tuning technique of multivariable Proportional-Integral-Derivative (PID) controllers applied to a neonatal incubator. The method consists of the identification of a first order plus dead-time (FOPDT) model of a two-input and two-output (TITO) process and the utilization of this model to retune the PID controller. To perform the identification, a sequence of setpoint step changes, one at each input, is applied until the steady-state is reached and the step responses are evaluated by analytic formulae. As a TITO process, the temperature and relative humidity control on a neonatal incubator is a proper application for this new method. Experimental results are presented and prove the effectiveness of the method.