NETWORKED ROBUST PREDICTIVE CONTROL SYSTEMS DESIGN WITH PACKET LOSS

Quang T. Nguyen – Vojtech Veselý – Alena Kozáková – Pavel Pakshin

   The paper addresses problem of designing a robust output feedback model predictive control for uncertain linear systems over networks with packet-loss. The packet-loss process is arbitrary and bounded by the control horizon of model predictive control. Networked predictive control systems with packet loss are modeled as switched linear systems. This enables us to apply the theory of switched systems to establish the stability condition. The stabilizing controller design is based on sufficient robust stability conditions formulated as a solution of bilinear matrix inequality. Finally, a benchmark numerical example-double integrator is given to illustrate the effectiveness of the proposed method.

Keywords: model predictive control (MPC), networked control systems (NCSs), packet-loss, parameter dependent quadratic stability (PDQS), bilinear matrix inequality (BMI)

[full-paper]