Iterative Feedback Tuning of Model-Free Controllers

Iterative Feedback Tuning (IFT) and Model Free Control (MFC) are strategies from the Data Driven Control (DDC) class. The first is a controller tuning method and the second is a control algorithm, both based only on the data collected from the process. In this paper, a new strategy that improves the MFC control law performance with the IFT algorithm is proposed. To tune the MFC controller parameters, several signals are collected from the process, and an iterative procedure is started. The objective of the tuning procedure is to minimize a criterion that leads to a new set of parameters for the MFC algorithm. The proposed algorithm is tested in a speed control application on an internal combustion engine with a complex model. The results obtained with the intelligent P(iP)-IFT algorithm are compared with PI and iP control laws to analyze the impact of IFT on the iP control loop.