Centralized model predictive control of autonomous driving vehicles with Lyapunov stability

The purpose of this paper is the development of a centralized model predictive control strategy for vehicle dynamics with focus on autonomous driving vehicles. The employed predictive control is multivariable, controlling the steering maneuvers and longitudinal speed. The vehicle speed is controlled based on the powertrain dynamics, while the steering maneuvers are controlled based on the vehicle lateral dynamics. The proposed algorithm makes use of linear models of vehicle and powertrain in the controller design, and for reliable results, simulation are carried out by considering a nonlinear vehicle model as the plant. For ensuring system stability, control Lyapunov functions are also included in the optimization algorithm. The obtained simulation results show that the proposed multivariable predictive control approach could provide very good performances in practical use.