Modeling and Simulation of Mixed Vehicles and Bicycles Flow on Urban Road Segments

Mixed vehicle and bicycle flow is the most common phenomenon on urban road segments with traffic markings. Existing simulation models cannot consider the behavioral production process and simultaneously model two-dimensional behaviors. This paper proposes a two-dimensional model based on a three-layer “perception-decision-action” framework. The perception layer selects the dominant interactive object from the multi-interaction surroundings. The decision layer develops a rule-based criterion to select an acceptable behavior in different alternatives. The plan layer provides an acceptable trajectory using the Bezier-curve-based method and the action layer stores a series of operational models to specialize the selected behavior into accelerations. The proposed model is tested by reproducing 185 vehicle trajectories and 448 bicycle trajectories on a road segment in Shanghai. The simulation reproduces the variation of trajectories, while the coverage rates are 85.6% and 84.1%, respectively. Travel time distributions of empirical and simulation data show no significant difference.