A Flexible Strategy for Efficient Merging Maneuvers of Connected Automated Vehicles

Abstract

Merging is a challenging task for automated vehicles. This paper proposes a strategy for connected automated vehicles (CAVs) to guide merging on-ramp vehicles efficiently while ensuring safe interactions with the mainline vehicles. Point-mass kinematic models are used to describe 2-D vehicle motion and receding horizon control is used to generate optimal trajectories of interacting vehicles. The strategy determines the optimal merging time instant for merging vehicles and acceleration of all involved vehicles to minimize deviation from the preceding vehicles' speed, deviation from preferred inter-vehicle gaps, accelerations, and the time spent merging. The strategy builds on a pre-determined order of vehicles passing the conflict zone but is not restricted to fixed merging points as previous research assumes. It resembles human-like behavior in the sense that on-ramp vehicles will accept smaller gaps when approaching the end of the acceleration lane. The performance of the strategy is demonstrated in simulations.