Hindrance-Aware Platoon Formation for Connected Vehicles in Mixed Traffic

Abstract

Aiming at traffic efficiency improvement, vehicle platoon becomes an essential driving technique to control the synchronous driving of connected and automated vehicles. Since the mature automated driving technology and cooperative willingness cannot be ensured for all vehicles, a mixed traffic paradigm that incorporates both cooperative vehicles (CVs) and non-cooperative vehicles (NCVs) will be inevitable. Due to the traffic hindrance and intermittent connections caused by the randomly distributed NCVs, traffic efficiency would be limited in the mixed traffic. Therefore, a hindrance-aware platoon formation scheme is proposed for the mixed traffic in this paper, which aims at coordinating CVs to maximize overall traffic velocity while minimize collision risk and fuel consumption. Specifically, a vehicle connection model is first established to construct the hindrance-aware relationships among vehicles, based on which, the intermittent connections of CVs are formulated as a mixed integer nonlinear programming (MINLP) problem. To reduce the search space, a novel cross-vehicle matrix is designed to represent the hindrances that each CV faces, then a traffic-maneuver-tree construction algorithm is proposed to search for feasible and hindrance-aware platoon formation connections. A length-dynamic inverse influence zone is constructed, then two decision indicators of platoon leader and follower are decoupled by evaluating leadership qualifications of all CVs. Given the hindrance-aware connections and leadership qualifications, a hindrance-aware platoon formation algorithm is proposed to jointly optimize the speed, safety and energy efficiency of CVs. To verify the effectiveness of the platoon formation mechanism, extensive simulations are conducted, and the results reveal that the road utilization, traffic capacity improvement and safety assurance in mixed traffic can be enhanced.