Abstract
Improving road safety requires not only reacting to accidents but also identifying and addressing accident risk before crashes occur. This study investigates harsh braking maneuvers - instances of abrupt vehicle deceleration - as leading indicators of potential safety issues in urban traffic. Such events often reflect sudden conflicts or unexpected changes in the driving environment and can serve as a valuable proxy for assessing road safety conditions. Using crowdsourced vehicle data from the urban area of Braunschweig, we perform a large-scale analysis of harsh braking events at road-link level. Rates of braking events are normalized and smoothed to account for exposure, and their spatial distribution is compared with accident occurrence. The results reveal strong correlations between braking activity and accidents, but also highlight locations where braking and accident rates diverge. These include cases where braking occurs frequently without corresponding accident records, suggesting latent risks or the influence of infrastructure, which is adapted to harsh braking maneuvers. The findings demonstrate that harsh braking data can complement traditional accident statistics and enable a proactive approach to safety analysis. By identifying potentially hazardous locations before accidents occur, this method supports data-driven traffic planning and contributes to the broader goal of reducing road casualties.

This publication has 13 references indexed in Scilit: