Model predictive control for connected and autonomous vehicles at road intersections

Abstract

This work presents a model predictive control (MPC) to coordinate connected and autonomous vehicles (CAVs) with a Vehicle-to-Vehicle (V2V) communication when they enter intersections. With the purpose of minimizing energy as well as passing the intersection smoothly, it uses an individual linear quadratic optimal controller for each CAV, with a predefined path, that will respect mixed-integer linear constraints to guarantee collision avoidance in relation to the nearby vehicles. This method solves different scenarios with a different number of CAVs crossing the intersection, coming from more than one road, including a platoon formation. The results show that MPC is an efficient technique to integrate multiple CAVs to collaborate with the mutual objective of join merging zones without accidents.