Sensitivity of road safety indicators in normal and crash cases

Road traffic safety is receiving increasing attention. Different real-time safety indicators relying on traffic sensor data can be found in the literature. Some are based on the posterior analysis of crash and traffic data; while others are based on real time traffic risk assessment using microscopic parameters (e.g., short time headways). Other non-traffic related parameters have an impact on safety too, such as day/night, state of the road (dry/wet), etc. Yet, in most studies, not all parameters are available or considered. In this paper, individual traffic, road weather and crash data are considered on a Swiss canton (Vaud). All data sources are available for at least 4 years (2002-2005) on more than 200 km of motorways. This comprehensive data enables the authors to study 1) the adaptability to Vaud canton of different safety indicators from the literature, and 2) the sensitivity of the safety indicators on Vaud canton data, both under normal situations and crash situations and specially the rear-end crashes.