Adaptive control with moving actuators at motorway bottlenecks with connected and automated vehicles

Abstract

Connected and automated vehicles (CAVs) have the potential to improve the operation of future road traffic systems. In this paper, we propose a control method that uses CAVs as dynamic actuators to improve the capacity at motorway bottlenecks robustly. The proposed approach has been designed for mixed traffic flow using the fundamental diagram model of mixed traffic flow as a control activation tool. In order to implement our approach, we assume that the availability of detectors at motorway are able to obtain the density in real-time. The idea is that assuming a certain percentage of CAVs presence on the road, such vehicles can be used as mobile actuators to perform speed coordination tasks. Furthermore, the aim is to transfer the delays observed at the bottlenecks upstream on the motorway, where the conditions are more homogeneous. The proposed approached can be generalized and used in bottleneck scenarios with or without additional inflow from an on-ramp. According to the designed control strategy, when the traffic density at the bottleneck satisfies the activation condition, the CAVs will shift to moving actuator mode and generate a new speed profile. The objective is to improve traffic flow at the downstream bottleneck and also smooth the upstream arrival vehicle speed, thus improving the overall throughput of the network. The method has been evaluated through microscopic simulation experiments conducted with scenarios on the real-case study, a motorway in Antwerp, Belgium. The results show significant improvements in reducing traffic density and improving travel speed both locally at the control area, as well as at network level. Comparative experiments under different penetration rates show that the proposed method remains robust to the percentage of CAVs on the road. Finally, it can significantly reduce vehicle delays and prevent over-congested conditions, and also improve the traffic flow rate, even for relatively low penetration rates of CAVs.