Analysis of Vehicle Sequencing at Smart Intersections: A PDMP Approach

With the development of connected and autonomous vehicles (CAVs), signal-free intersections with both efficiency and safety is under great need. However, macroscopic performance metrics of smart intersections have not been well evaluated. In this paper, we model a signal-free intersection as a piecewise-deterministic Markov process. We use Foster-Lyapunov theory to analyze the macroscopic properties such as intersection capacity and waiting time of the system, and compare the performance of three typical sequencing policies: first-in-first-out (FIFO), minimal switch-over (MSO) and longer queue first (LQF). We reach the conclusion that MSO performs the best while LQF performs the worst. Finally, we apply the results to a numerical example to gain insights for sequencing policy design and validate the proposed model using simulation of urban mobility.