Iterative Distributed Model Predictive Control in the Presence of Coupling State Constraints.

In this paper, a distributed model predictive controller (MPC) is developed for the control of distributed dynamic systems with bounded dynamic couplings subject to state constraints, coupling state constraints and input constraints. For linear dynamics, dynamic couplings can be bounded due to coupling constraints. In the proposed control scheme, all subsystems solve their respective local optimization problem in parallel requiring only neighbor-to-neighbor communication. Consistency constraints, which ensure that a subsystem's actual state trajectory remains in the neighborhood of its previously communicated reference trajectory, are employed to bound the uncertainties resulting from the simultaneous evaluation of the local optimization problems. The disturbances caused by these uncertainties are handled with a robust tube-based MPC approach. Due to the bounded dynamic couplings among the subsystems, an iterative procedure can be invoked to overcome the restrictions on the degrees of freedom of the local optimization problems caused by the consistency constraints. In the end, the controller design procedure is illustrated with an analytically tractable example, and the algorithm's applicability is demonstrated with a simulation.