Behavioral investigation of stochastic lateral wandering patterns in mixed traffic flow

The vehicles’ lateral movement exhibits inherent stochastic behavior, and when considered collectively, the traffic flow demonstrates certain lateral wandering pattern which serves a fundamental input for pavement design and sustainability analysis. Naturalistic driving trajectories show that the vehicles on neighboring lanes tend to interact laterally, which leads to correlated wandering on a multi-lane road. Compared with human-driven vehicles (HDV), autonomous vehicles (AVs) have a more centered wandering pattern, which accelerates the pavement fatigue. As the market penetration rate of the AVs changes, the wandering pattern of the mixed AVs flow would evolve accordingly, which further influence pavement performances. The evolutionary wandering pattern of mixed AVs flow which considers the interactions between neighboring lanes’ traffic flow is not fully studied yet. This research fills this gap with a stochastic behavioral lateral movement model that bridges the microscopic behavior and lateral wandering pattern. Firstly, a lateral movement model is developed, which assumes that lateral dynamics are influenced by lane markings, road boundaries, intra-lane interaction, and noise. The model is expressed in a stochastic differential form. The solution of the model which gives the lateral wandering is expressed by the Euler-Maruyama scheme, Fokker-Planck equation and Markov chain. Then the model is employed to investigate the wandering patterns under different scenarios. The results show that the model's outputs are consistent with current research results and empirical observations. This research provides additional insights regarding lateral wandering under mixed autonomous vehicle flow, and offers a theoretical tool for future road pavement design.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01894336
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Sep 25 2023 2:46PM