Investigation of the Impact on Throughput of Connected Autonomous Vehicles with Headway Based on the Leading Vehicle Type

Abstract

Although the future era of autonomous driving is seen as a solution for many of the current problems in traffic; the introductory phase, with low penetration rates of connected-autonomous vehicles (CAVs), might lead to lower capacities. This forecast is based on certain assumptions that the CAVs can operate more efficiently when communicating and cooperating—already proved in real tests—therefore in practice, they can keep smaller following headways. However, it is envisioned that they might keep larger headways to other conventional vehicles for safety reasons. Lower connected-autonomous vehicle (CAV) penetration rates lead to a reduction in the overall vehicle throughput, then with increasing penetration rates, throughput is recovered and eventually improved. Simulations demonstrate that the impact on vehicle throughput depends on the car following headway and penetration rate. Based on this potential reduction in the maximum throughput for low penetration rates, the aim of this paper is the mitigation of this phenomenon at urban intersections through a possible managing solution to sort CAVs and a pre-set green-time start. A microsimulation model has been calibrated using PTV Vissim to reflect this operating solution, using new possibilities as leading vehicle class dependent headway settings and formula-based routing for sorting vehicles at a two-lane intersection entry. This approach allows the formation of platoons at intersections and uses their effectiveness even at low CAV penetration rates. The tested scenario is simplified to through traffic without turnings maneuvers and the results show that the potential loss in throughput is canceled and reductions in the control delay can reach 17% for oversaturated conditions.