Robust and efficient data-driven predictive control

TL;DR

This paper introduces a robust, data-driven predictive control method that is more sample-efficient and computationally efficient, using noisy input-output data to ensure stability and practical performance in linear systems.

Contribution

It presents a novel, noise-robust data-driven predictive control scheme that requires less data and computation than existing methods, with proven stability and practical applicability.

Findings

The scheme is more sample-efficient than existing methods.

It guarantees recursive feasibility and practical stability.

Performance is validated on a four tank system case study.

Abstract

We propose a robust and efficient data-driven predictive control (eDDPC) scheme which is more sample efficient (requires less offline data) compared to existing schemes, and is also computationally efficient. This is done by leveraging an alternative data-based representation of the trajectories of linear time-invariant (LTI) systems. The proposed scheme relies only on using (short and potentially irregularly measured) noisy input-output data, the amount of which is independent of the prediction horizon. To account for measurement noise, we provide a novel result that quantifies the uncertainty between the true (unknown) restricted behavior of the system and the estimated one from noisy data. Furthermore, we show that the robust eDDPC scheme is recursively feasible and that the resulting closed-loop system is practically stable. Finally, we compare the performance of this scheme to existing ones on a case study of a four tank system.