Optimal lane management policy for connected automated vehicles in mixed traffic flow

Abstract

The management of lanes for mixed traffic flow consisting of connected automated vehicles (CAVs) and human-driven vehicles (HDVs) can effectively improve the operational efficiency of highways. This paper proposes an optimal lane management policy for connected automated vehicles in mixed traffic flow on highways. Firstly, the functioning characteristics of the mixed traffic flow are analyzed in three aspects, including car-following mode, headway, and platoon size. Secondly, the vehicle probability distributions under different lane management strategies are derived based on the Markov chain and conditional probability. Then, a calculation model for the capacity of mixed traffic flow was constructed by considering the features of lane management strategies, and sensitivity analyses were performed on the model parameters. Finally, a joint optimization model of CAVs dedicated lanes is proposed with the objective of maximum throughput based on constraints such as lane operation strategies, vehicle type ratios, and road resources. The numerical experiments lead to the following conclusions. (1) For two-lane highways, when the traffic demand is less than 6000 veh/h, the mixed driving lane management strategy should be applied; and when the traffic demand reaches 8000 veh/h, and the CAVs penetration rate is more than 0.56, one CAVs dedicated lane is required. Also, the higher the traffic demand, the lower the critical penetration rate of CAVs. (2) In the three-lane highway scenario, no dedicated lane is required when traffic demand is less than 8000 veh/h; one CAVs dedicated lane is needed when demand exceeds 8000 veh/h with CAVs penetration rate greater than 0.38; and two CAVs dedicated lanes are necessary when demand increases to 12800 veh/h and CAVs penetration rate is more than 0.79. (3) As for four-lane highways, the minimum traffic demand for installing one dedicated lane is 15600 veh/h, and the critical penetration rates for one, two, and three dedicated lanes are 0.32, 0.71, and 0.96, respectively. (4) The increase of CAVs dedicated lane usage willingness and platoon size promote the throughput, and the former CAVs dedicated lane usage willingness works more obviously. These research conclusions can provide theoretical and applied practical support for CAVs dedicated lanes management in the future.