Data-driven design of explicit predictive controllers using model-based priors

TL;DR

This paper introduces a data-driven method to directly design explicit predictive controllers by leveraging model-based priors, eliminating the need for system identification and enabling stability verification before deployment.

Contribution

It presents a novel approach that optimizes explicit predictive controllers directly from data using known PWA structure, incorporating stability checks and noise handling techniques.

Findings

Effective design of explicit predictive controllers from data

Automatic closed-loop system modeling and stability verification

Robustness to noise through regularization and averaging

Abstract

In this paper, we propose a data-driven approach to derive explicit predictive control laws, without requiring any intermediate identification step. The keystone of the presented strategy is the exploitation of available priors on the control law, coming from model-based analysis. Specifically, by leveraging on the knowledge that the optimal predictive controller is expressed as a piecewise affine (PWA) law, we directly optimize the parameters of such an analytical controller from data, instead of running an on-line optimization problem. As the proposed method allows us to automatically retrieve also a model of the closed-loop system, we show that we can apply model-based techniques to perform a stability check prior to the controller deployment. The effectiveness of the proposed strategy is assessed on two benchmark simulation examples, through which we also discuss the use of regularization and its combination with averaging techniques to handle the presence of noise.