Distributed Model Predictive Control for Automated Intersection with Multiple Maneuvers

Intersections play a vital role in traffic management. Vehicle-to-Vehicle (V2V) communication brings breakthroughs to the safety and efficiency of intersections. In this paper, a distributed model predictive control approach is proposed which enables multiple vehicles cross the intersection simultaneously without collision. A simplified vehicle dynamic model with longitudinal and lateral motions coupled is utilized in local controller to describe the dynamic properties of the vehicle. A scenario where several vehicles coordinate to cross an intersection accounting for straight and turning maneuver is considered. In this case, the motions generated by vehicles locally in time are relied on optimization of the problem, incorporating constraints based on collision avoidance and local vehicle dynamic. Finally, the effectiveness of the approach is verified in CarSim and MATLAB co- simulation.