Calibration of an interrupted traffic flow system using NGSIM trajectory data sets

This paper calibrates an interrupted traffic flow system model using the trajectory datasets provided by Next Generation SIMulation (NGSIM) program. The realistic lane changing may cut off the traffic flow into discontinuous flow, giving rise to congestion and vehicle delay. The aim of this paper is to study the car-following behaviors of drivers in real traffic scenes in which lane changing happens. The interrupted traffic flow system model is calibrated by minimizing the deviations between the observed dynamics and the simulated trajectory with a nonlinear optimization procedure-Levenberg-Marquardt Algorithm. The reliability and robustness of the nonlinear fits are assessed for the deviation from one trajectory, acceleration and deceleration process respectively. The required errors are 0.0172 and 0.2845, which are sameness to representative one obtained in previous studies. Therefore, a used widely model that reflects the lane changing situation in reality is achieved.