Abstract
The study investigates the impacts of isolated and dedicated lanes for autonomous vehicles on freeway traffic flow, focusing on mixed traffic scenarios where autonomous vehicles coexist with human-driven vehicles. Given the anticipated long-term coexistence and the complex dynamics of mixed traffic, dedicated lanes for autonomous vehicles have been explored globally, but their benefits remain unclear. This research constructs detailed models of real-world freeway scenarios, calibrated with parameters from car-following models tailored for autonomous driving. By comparing traffic flow under various dedicated lane policies with mixed flow environments, the study reveals nuanced effects on overall traffic characteristics, particularly as the market penetration rate of autonomous vehicles increases. The findings show that lane policies significantly influence traffic throughput and efficiency, with certain policies demonstrating clear advantages at higher autonomous vehicles penetration levels. Results show that bidirectional and unidirectional isolation policies enhance traffic efficiency at higher market penetration rate (above 70%), while less restrictive policies are more effective at lower market penetration rate (below 50%). These findings emphasize the need to tailor lane management strategies to autonomous vehicles' market penetration rate, offering valuable insights for optimizing traffic flow and easing the transition to mixed traffic environments.