Event

Agent framework discovers car-following patterns from trajectory data: the AERM model

1 reports · 1 sources

Jiamei Xu et al. propose an agent-centric framework that integrates physical constraints to empirically discover interpretable car-following patterns from multi-source trajectory data, yielding the Adaptive Equilibrium Response Model (AERM), which couples a state-dependent relative-velocity response with a spacing-dependent equilibrium-speed recovery. The framework discovers AERM on NGSIM US-101, highD and exiD data, and compares it with seven classical car-following models on Shanghai naturalistic driving data, NGSIM I-80 and Waymo trajectories: under per-trajectory calibration, AERM has the lowest spacing and speed reconstruction errors; under dataset-level shared parameters it remains competitive.