Robust offset-free nonlinear model predictive control for systems learned by neural nonlinear autoregressive exogenous models

TL;DR

This paper develops a robust offset-free nonlinear model predictive control method using neural network models with stability guarantees, ensuring accurate tracking despite disturbances and model uncertainties.

Contribution

It introduces a novel tube-based MPC scheme that guarantees offset-free tracking for NNARX models with stability properties, addressing robustness and stability in nonlinear control.

Findings

Effective offset-free tracking demonstrated on water heating system

Robust stability maintained under model-plant mismatch

Numerical simulations confirm control scheme's effectiveness

Abstract

This paper presents a robust Model Predictive Control (MPC) scheme that provides offset-free setpoint tracking for systems described by Neural Nonlinear AutoRegressive eXogenous (NNARX) models. The NNARX model learns the dynamics of the plant from input-output data, and during the training the Incremental Input-to-State Stability (${\delta}$ISS) property is forced to guarantee stability. The trained NNARX model is then augmented with an explicit integral action on the output tracking error, which allows the control scheme to enjoy offset-free tracking ability. A tube-based MPC is finally designed, leveraging the unique structure of the model, to ensure robust stability and robust asymptotic zero error regulation for constant reference signals in the presence of model-plant mismatch or unknown disturbances. Numerical simulations on a water heating system show the effectiveness of the proposed control algorithm.