The development of freeway crash prediction models using intelligent transportation systems (ITS) archived data could be a substantial advancement in the field of real-time traffic management. Such models not only are expected to improve safety but also may go a long way to improve freeway operations by reducing incident-related congestion.
Because there is a need to use real-time traffic data emanating from loop detectors, the approach differs distinctly from previous studies estimating crash frequencies or rates on a certain freeway section through aggregate measures of flow (such as average daily traffic or hourly volumes).
Although the authors try to establish a relationship between the patterns in precrash data from detectors surrounding the crash location, it is imperative that the time of the historical crashes is known with precision.
This feature proposes a shockwave and rule-based methodology to estimate the time of the crash and then identifies how much time and distance ahead of crash occurrence loop data may be used to predict the impending hazard. The final objective is to predict the possibility of crashes on freeways using real-time loop data.
Available at: http://works.bepress.com/apande/9/