Modelling the road network capacity considering residual queues and connected automated vehicles

Abstract

This paper investigates the network capacity considering residual queues and connected automated vehicles. First, the weighted network outflow is introduced to measure the network capacity. Then, two bilevel models are built to calculate the network capacity with residual queues under the user equilibrium (UE) principle and the system optimum (SO) principle, respectively. The UE-based network capacity model is employed to estimate the capacity of the dedicated road network for human-driven vehicles (HVs), while the SO-based network capacity model is employed to estimate the capacity of the dedicated road network for connected automated vehicles (CAVs). Furthermore, by combining the two proposed network capacity models, a model to allocate link capacity between CAV lanes and HV lanes to maximise the whole network capacity is presented, in which the CAV travellers’choice behaviour of driving modes is considered. Finally, through two numerical examples, the characteristics of the proposed models are demonstrated and their potential applications are discussed. The results show that the proposed models are capable of 1) identifying critical links and potential critical links of networks, 2) guiding the land-use development and macroscopic travel demand management, 3) evaluating network optimisation schemes, and 4) allocating the road supply to CAVs and HVs efficiently.