Development and calibration of the Anisotropic Mesoscopic Simulation model for generalized vehicular traffic modeling
The development and deployment of Intelligent Transportation Systems (ITS) are heavily dependent on the availability of timely and accurate wide-area estimates of prevailing and emerging traffic conditions and need a mesoscopic traffic simulation model which is accurate in capturing traffic dynamics and efficient computationally. However, there does not exist such a simulation model which integrates anisotropic property of traffic flow. This research proposes an Anisotropic Mesoscopic Simulation (AMS) concept in which the speed of a vehicle is affected only by the density of its Speed Influencing Region (SIR) which is defined as the vehicle's immediate downstream roadway section in which the stimulus is considered significant enough to influence the vehicle's speed response. This research develops an AMS model for homogeneous highways and extends it to heterogeneous highways. AMS model is a vehicle-based mesoscopic traffic simulation model that explicitly considers car-following mechanism to update vehicle speed based on macroscopic speed-density relationships. This research also develops an AMS calibration methodology in which a least squares optimization formulation is adopted. U.S. Interstate 80 dataset developed by Next Generation Simulation (NGSIM) program of Federal Highway Administration (FHWA) is used to calibrate these models. The calibration results indicate that 250 feet is the optimal SIR length for homogeneous highways. The validity of calibration results is validated through an AMS simulation program. This research also generalizes AMS model to merging traffic. AMS models for major merge areas and ramp-freeway merge junctions are developed. U.S Highway 101 dataset developed by NGSIM program is used to calibrate these models. Simulation results indicate that AMS model for major merge areas has satisfactory flow distribution property, which is consistent with Jin and Zhang's demand-based distribution scheme.
Song, Houbing, "Development and calibration of the Anisotropic Mesoscopic Simulation model for generalized vehicular traffic modeling" (2006). ETD Collection for University of Texas, El Paso. AAI1439478.