Distributed Data-Driven Control of Transportation Networks

This paper presents a data-driven control for a network of autonomous vehicles. The general scheme of cruise control is implemented for a set of vehicles defined by a nonlinear equation. Koopman operator allows representing the nonlinear system in the Koopman space (lifted space), suitable to get a linear observer that is used for distributed control instead of the nonlinear model. The proposed control simplified the control design and also allows to use a linear Model Predictive Control (MPC) in a distributed form.