Traffic automation and lane management for communicant, autonomous, and human-driven vehicles

The advent of autonomous driving technologies has created a crucial need for upgrading conventional traffic control and lane management strategies in large cities. In this research, we design an optimal lane management strategy for corridors with a heterogeneous demand of human-driven, autonomous, and communicant autonomous vehicles (HVs, AVs, and CAVs). In a monocentric city setting, we dynamically control the inflow of the network by optimizing the size of CAV platoons in the corridors based on the instantaneous condition of the integrated system. These corridors can potentially have three types of lanes for vehicles with different levels of automation technology. We model the multiple lane type corridors as sets of parallel bottlenecks with general distributions of multiclass demand. The dynamics of the congestion in the network is also modeled using the macroscopic network fundamental diagram (MNFD). To study the impacts of the rise in the penetration rate of AVs and CAVs on the performance of the system, we derive a closed-form representation of the model. We show that the increase of the delay in the network with the rise in the penetration rate of AVs and CAVs can have a stable, an unstable, or a hybrid pattern. To optimize the system, we minimize a weighted summation of the experienced delay in the corridors and the total travel time in the urban network by optimizing the number of lanes of each type and the dynamic size of the CAV platoons. The results of the San Francisco case study show that implementing an optimal lane management strategy can reduce the experienced delay in the corridors up to 78% with a rise in the AV/CAV penetration rate. By dynamically controlling the size of the CAV platoons in the automated highway of the Bay Bridge, we limit the increase of the travel time in the downtown network as low as 5%.