CAV-dedicated lane sharing strategy of expressway in mixed traffic environment: A novel dynamic lane-level tolling method

In a mixed network environment on expressways, designating dedicated lanes for connected and autonomous vehicles (CAVs) can effectively enhance road capacity. However, at low CAV penetration rates, setting dedicated lanes may cause inefficient use of road resources and potential congestion. To improve the utilization efficiency of dedicated lanes, this paper develops a dynamic tolling strategy for connected human-driven vehicles (CHVs) to use CAV-dedicated lanes through a pay-and-share mechanism. This dynamic tolling strategy involves two primary components. First, an optimized lane-level cell transmission model (CTM) is developed to incorporate the proportion of CHVs willing to pay for CAV-dedicated lanes access, combined with a user equilibrium (UE) model based on heterogeneous time values to allocate traffic flow. Second, a model predictive control tolling (MPCT) strategy is proposed, using the refined lane-level CTM as the predictive model and the dung beetle optimizer (DBO) algorithm to calculate the optimal toll rate sequence. Finally, the effectiveness of the proposed MPCT strategy is evaluated through a comparative analysis against with a feedback control tolling (FCT) strategy and a no control (NC) baseline. Simulation results indicate that the proposed MPCT strategy can significantly improve traffic efficiency and safety while also contributing to reduced fuel consumption.