Answers are grounded in this paper and cite their sources.
Observational study uncovers behavioral and environmental triggers of traffic violations among commuting e-bike riders, highlighting cumulative microbehavioral risks and a 4-meter lane threshold.
To uncover the causal mechanism underlying e-bike violations during weekday commuting peak hours, this study develops an e-bike violation-behavior dataset using video recordings collected at signalized intersections. A composite violation risk indicator is then proposed. Subsequently, an interpretable Bayesian-optimized light gradient boosting machine (BO-LightGBM) model integrated with bivariate partial dependence plots (PDP) is developed to identify key influencing factors and reveal the interaction effects. The results indicate that violation risk is primarily driven by behavioral characteristics, particularly two-stage crossing behavior, observable looking behavior, and following behavior. Interaction analyses further reveal complex nonlinear patterns across factors. First, the occurrence of two-stage crossing, observable looking, and following behaviors markedly increases violation risk, and their co-occurrence exhibits clear cumulative effects. Second, violation risk is highest for riders traveling in groups of 7 to 10 and under moderate motor vehicle volumes (60 to 80 vehicles/5 min). In contrast, low (0 to 1 m/s) and moderate (2.5 to 3.5 m/s) riding speeds correspond to distinct decision patterns that are consistent with opportunistic waiting to violate and accelerated rushing-through, respectively. Finally, nonmotorized lane width shows a safety threshold of approximately 4-m widths below this threshold increase the risk of violation by 12%.
No takes yet. Share an insight, caveat, or question.
Wen et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: