Distributed switched model-based predictive control for distributed large-scale systems with switched topology

TL;DR

This paper introduces a distributed robust switched model predictive control method for large-scale systems with dynamically changing topologies, ensuring stability and constraint satisfaction despite unknown switching signals.

Contribution

It proposes a novel distributed control scheme that handles unknown switching signals and topology changes using a robust tube-based approach with invariant sets.

Findings

The proposed DSwMPC guarantees stability and constraint satisfaction.

Simulation results validate the effectiveness of DSwMPC under various switching scenarios.

DSwMPC outperforms centralized and decentralized approaches in response quality.

Abstract

Distributed switched large-scale systems are composed by dynamically coupled subsystems, in which interactions among subsystems vary over time according to an exogenous input signal named switching signal. In this paper, we present a distributed robust switched model-based predictive control (DSwMPC) to control such systems that are subject to local state and input constraints. The proposed method guarantees stabilizing the origin of the whole closed-loop system and ensures the constraints satisfaction in the presence of an unknown switching signal. In the distributed model-based predictive control (DMPC) used in this work, by considering the interactions among subsystems as an additive disturbance, the effect of the switch is reflected on the dynamic equation, local, and consistency constraint sets of the nominal subsystems. In the DSwMPC, to compensate the effect of switching signal which creates a time-varying network topology, a robust tube-based switched model-based predictive control (RSwMPC) with switched robust control invariant (switchRCI) set as the target set robust to unknown mode switching is used as local controller. The scheme performance is assessed using three typical examples. In the first example, the switching times are unknown in prior, but the next neighborhood sets are assumed to be known in prior. In the second and third cases, both of them are supposed to be unknown in prior. The simulation results show that the input and state constraints are satisfied by the proposed DSwMPC at all times. They also validate that the closed-loop system converges to the origin. Also, a comparison of the DSwMPC with a centralized SwMPC (CSwMPC) and a decentralized SwMPC (DeSwMPC) shows that the DSwMPC outperforms the DecSwMPC and also the shapes of response curves under the CSwMPC are very similar to those obtained by the DSwMPC.