Network screening for specific collision types at urban signalized intersections - conventional and spatial methods

Transportation authorities are always looking for ways to improve road safety since vehicle collisions cost Canada 25 billion dollars of capital loss and around 2800 deaths each year. An important step in improving road safety is to sieve out the problem sites through network screening processes. screening for specific accident types is discussed in this thesis, using signalized intersections in Toronto in an illustrative application. each such collision type is associated with corresponding countermeasures, which allows the engineer to rank the entities with specific remedies in mind. In this way, the effectiveness of road network screening can improved through targeted treatments. Three different screening methods are introduced and compared; procedures for selecting entities from screening results by different methods are also presented. A process for ranking jurisdictions by regions is proposed. This is a method which combines the conventional network screening techniques with geographic information system (GIS) tools. The GIS can integrate the spatial information of a selected area with the conventional accident and road characteristic data and facilitate network screening by region.