A Cooperative Control Strategy of Connected and Autonomous Vehicles in Ramp Merging Areas with Mixed-Autonomy Traffic

Considering mixed-autonomy traffic with connected and autonomous vehicles (CAVs) and human-driven vehicles (HVs), a CAV cooperative control strategy was proposed to improve the traffic flow efficiency in freeway ramp merging areas. The roadside unit controlled the movements of CAVs both on the ramp and on the mainline before the merging point, according to real-time traffic conditions. Microscopic simulations based on cellular automata were performed to evaluate the influences of the proposed cooperative control strategy on the average vehicle delays of the mainline, the ramp, and the entire merging area. Sensitivity analyses were performed with varying CAV market penetration rates and traffic demand levels. Results show that, compared to no-control case, the proposed cooperative control strategy may effectively reduce the average vehicle delay of the mainline. Though the average vehicle delay of the ramp increases, the average vehicle delay of entire merging area is still improved, because the mainline flow volume is usually much larger than the ramp flow volume. Furthermore, the proposed cooperative control strategy only requires a limited number of control variables, making it feasible to determine the optimal control variables based on sensitivity analysis.

I. Introduction

With the growth of CAV market penetration rate, CAVs and HVs will coexist in the transportation system for a long time. Many works have been carried out to investigate the impacts of mixed-autonomy traffic on, e.g., the road capacity [1–3] and the energy consumption [4]. CAVs can drive under control according to external instructions. Therefore, many studies proposed to improve the traffic flow propagation in different scenarios via controlling the movements of CAVs [5–8]. One portion of studies have tried to apply cooperative control of CAVs to urban intersection control [9–11] and another portion of studies investigate the freeway ramp merging problem [12–13]. Most of these works focus on pure automated traffic flow. And the works on cooperative control of CAVs in mixed-autonomy traffic is very limited. Recently, [14] proposed an optimization model to facilitate the cooperative decision-making for mixed-autonomy traffic in freeway merging areas. To derive the explicit model formulation, they need to assume that the trajectories of HVs are predictable in short-term. Such an assumption may bring safety issues when implemented in reality. Because any small error in the prediction of HV movements can lead to vehicle collision in the mixed-autonomy traffic.