Estimating Freeway Lane-Level Traffic State with Intelligent Connected Vehicles

Abstract

This paper proposes a methodology for estimating lane-level traffic state for freeways by fusing data from intelligent connected vehicles (ICVs) with fixed detector data (FDD) and probe vehicle data (PVD). With microscopic vehicle trajectories of ICVs and their surrounding vehicles, the proposed methodology integrates a multilane traffic flow model into the data assimilation framework based on extended Kalman filter (EKF), in which traffic measurement models are formulated for ICV data, PVD, and FDD, respectively, to fit their different characteristics. Simulation experiments are conducted to test the performance of the proposed methodology with various penetration rates of ICVs, using a set of simulated ICV data based on the Next Generation SIMulation (NGSIM) data sets. The results demonstrate that by utilizing only 3% to 5% ICVs in the mixed traffic, the proposed methodology could produce an accurate estimate of lane-level traffic speed and a reasonable estimate of lane-level traffic density.