Distributed multilane merging for connected autonomous vehicle platooning
Science China Information Sciences
In the context of coordination of connected autonomous vehicles (CAVs), the platooning operation is a promising application. The formulation of a single stream of CAVs is conducive to traffic efficiency and merging operations extend the benefits for multilane road users. However, the problem of simultaneous merging and platooning lacks comprehensive investigation. A solution is formulated in this paper through a new scheme that considers inter-vehicle safety distance constraints and distributed deployment utilizing local inter-vehicle information exchanges. A distributed consensus-based controller synthesized with a collision avoidance design is developed to direct the CAVs to maintain the velocity and spacing required to avoid inter-vehicle collisions. Furthermore, a framework fusing an agent motion model with vehicle controllers based on a dynamics model that facilitates both longitudinal and lateral controls is proposed, contributing to a cross-model planning-tracking controller. Theoretical proof of asymptotic stability of the proposed controller and its collision avoidance capability are also elaborated. The merging and platooning function was tested in a hardware-in-the-loop (HiL) experiment, demonstrating the precise tracking performance and comparable merging responses to a typical multiagent system. In comparison with trajectory-based merging algorithms, the proposed framework is able to achieve finer stepwise tracking results without centralized coordination or predefined trajectories.
Open Access Status
This publication is not available as open access
National Natural Science Foundation of China