Guaranteed Estimation for Distributed Networked Control Systems

This paper proposes a distributed estimation and control scheme for discrete linear time-invariant systems. The plant is monitored and controlled by a network of agents that collect information about the system evolution and apply control inputs to drive its behavior. Bounded disturbances and measurement noises are considered through a zonotope-based approach. It is proven that the state of the plant belongs to the prediction and estimation zonotopes constructed by all the agents at every time instant. The performance of the proposed technique is tested by simulation using a classical system: the inverted pendulum on a cart.