Dynamic capacity estimation of mixed traffic flows with application in adaptive traffic signal control

Abstract

Intersection control plays a vital role in influencing transportation efficiency inside urban areas. Connected and Automated Vehicle (CAV) technology enables frequent traffic information sharing through vehicular networks, which emerges as a promising way to reduce vehicle traveling time and improve intersection capacity. Meanwhile, with the gradual deployment of CAVs, the traditional traffic flow consisting of human driven vehicles will evolve into a mixed traffic flow composed of traffic participants with differing intelligence capabilities. Considering the evolution towards the mixed traffic environment, we propose a dynamic capacity-aware traffic signal control method: max-pressure for mixed traffic flow (MPMF) traffic signal controller. Firstly, the lane capacity is modeled and approximated based on the saturation flow rate in mixed traffic flow considering the penetration rate of CAVs. Then the dynamic capacity is involved in calculating pressure in the max-pressure method to indicate the importance of a path in mixed traffic flow. A modification strategy for the max-pressure input value is also proposed. An isolated intersection scenario was simulated first to assess the proposed method locally. A multi-intersection network experiment was also conducted to verify the network-level performance of the proposed MPMF method. Comparative results between the proposed MPMF method, the classic max-pressure control, and the existing fixed time control method demonstrate that the MPMF can effectively improve the performance of intersections and be suitable for the multi-intersection road network.