Promoting safe transit: analyzing bus accident patterns

Accident taxonomy is widely used by researchers and practitioners worldwide as a tool for understanding accident risks and designing effective policy measures to mitigate these risks. Interestingly, despite the usefulness of accident taxonomy for identifying accident risks and the growing interest in improving bus safety operations, information regarding the taxonomy of bus accidents is scarce. The current study provides a holistic perspective of the risk-factors underlying bus accidents by identifying prevailing bus accident typologies and evaluating their severity in the United States.In order to identify bus crash clusters based on their features, data from the General Estimates System (GES) crash database are clustered by means of a two-stage clustering method, consisting of self-organizing maps (SOM) followed by neural gas, Bayesian classification and unified distance matrix edge analysis. A multi-layer perceptron (MLP) neural network was employed to confirm the correctness and usefulness of the SOM-based clustering process.Five clusters are identified: (i) multi-vehicle collisions at intersections: vehicle encroaching or travelling; (ii) multi-vehicle collisions with school bus at intersection: distracted drivers; (iii) multi-vehicle collisions in road sections: infrastructure and traffic; (iv) single-vehicle bus accidents off-road: bus travelling and bus driver distraction at low speeds; (v) single-vehicle collisions with non-motorists: pedestrian and cyclists. The analysis points out conflicts among buses and other road users and indicates possible cluster-driven directions towards enhancing bus safety.