Microscopic Modeling of Non-Motorized Traffic Flow in Bicycle Lanes Based on a Modified Social Force Model Considering Non-Motorized Vehicles’ Behavioral Heterogeneity

In urban bicycle lanes, the variety of non-motorized vehicles (NMVs) and the discrepancies in riders’ cognitive and behavioral characteristics have resulted in the behavioral heterogeneity of NMVs in the mixed traffic flow, which has not been fully considered in existing microscopic simulation models of NMVs. To facilitate a more precise depiction of behaviors and interactions with other road users for NMVs with different patterns, this study proposed an agent-based social force model (SFM) to simulate the NMV flow in urban bicycle lanes. First, three types of behavioral patterns were identified from real-world NMV trajectories, including cautious, moderate and reckless. The perception and decision processes of NMV agents with different behavioral patterns were represented by multi-layer fuzzy logic models. Considering the NMV agent’s behavioral decisions of proceeding direction and belling intention, a modified SFM consisting of a driving force, repulsive forces and physical forces was constructed to model the NMV agent’s microscopic behaviors. The proposed model has been calibrated based on real-world NMV trajectories. The simulation results indicate that the proposed model could reproduce the operational characteristics of the NMV flow in bicycle lanes, and composition of NMVs with different behavioral patterns would have an impact on the microscopic interactions within the traffic flow.