Markov chain-based traffic analysis on platooning effect among mixed semi- and fully-autonomous vehicles in a freeway lane

Focusing on a freeway lane running with mixed semi-autonomous vehicles (semi-AVs) and fully-autonomous vehicles (fully-AVs), this study aims to investigate the heterogeneous platooning effect of the two types of autonomous vehicles (AVs) on the travel efficiency of the mixed traffic. Such impact in the mixed semi-/fully-AV traffic, to the best of our knowledge, has not been well addressed in the literature as the existing mixed traffic studies on platooning effect typically focused on the traffic scenes with only human-driven vehicles (HVs) and fully-AVs. The existence of semi-AVs, which are optionally controlled by human drivers and semi-autonomous driving systems, and their significant difference in platooning behavior from the fully-AVs have received very limited attention. To fill the gap, we first characterize the semi- and fully-AVs' difference in connection ability and willingness to platoon in the vehicle platooning process, and derive the heterogeneous platooning probabilities of semi- and fully-AVs. To formulate the platooned traffic after vehicle platooning process, we proceed to develop a novel Markov chain model to determine the probability distributions of vehicular platoon size and headway type among the mixed semi-/fully-AV traffic analytically. Furthermore, with the probability distributions, we derive the fundamental diagram (FD) and the capacity of the freeway lane for the mixed semi-/fully-AV traffic, and are thus able to investigate impact of platooning on the travel efficiency numerically. The developed models and unveiled insights in this study will help researchers and practitioners better understand the role, potential and prospect of automated transport system in the transitional stage with semi-AVs, which will be useful for guiding the mixed AV traffic management and control in near future.