Distributed Control and Estimation Scheme With Applications to Process Control

This brief proposes an innovative estimation and control scheme that enables the distributed monitoring and control of large-scale processes. The proposed approach considers a discrete linear time-invariant process controlled by a network of agents that may both, collect information about the evolution of the plant, and apply control actions to drive its behavior. The problem makes full sense when local observability/controllability is not assumed and the communication between agents can be exploited to reach system-wide goals. The design procedure guarantees system stability, allowing the designer to tradeoff performance and control effort. The obtained controllers and observers are implemented in a fully distributed fashion. To show the performance of the proposed technique, simulation results on a water delivery system are provided.