Wednesday, November 22, 2023, 11:59 - 12:45
Room 01-012, Georges-Köhler-Allee 102, Freiburg 79110, Germany
Distributed Model Predictive Control (MPC) addresses the control of cyber-physical systems while aiming to build on the success of centralized MPC. Real-time requirements challenge cooperative approaches with multiple optimizer iterations per control step, often enforce suboptimal control inputs in real-world implementations, and question closed-loop stability. We combine the real-time iteration framework from centralized nonlinear MPC with tailored decentralized numerical algorithms and present a real-time iteration framework for distributed MPC. This way, we obtain new stability guarantees for the distributed MPC of coupled nonlinear systems under suboptimal control inputs. Numerical examples and experimental results from robot formation control demonstrate the efficacy of the proposed approach.