Identification of highway segments with an unusually high occurrence of non-intersection crashes has been a particularly challenging problem for the Wisconsin Department of Transportation, because these crashes were scattered over approximately 9,500 miles of highway. The PRÈCIS floating highway segment algorithm was developed to identify crash rates and crash densities at any given point along the entire length of any undivided state trunk highway (STH). Once the entire set of STH mileage is processed through the algorithm, crash rates and crash densities can be processed to identify locations where these metrics exceed a predetermined value. Locations can be rankordered for treatment across the entire STH system. Crash rate information is then integrated on GIS maps and a graph providing visual feedback about higher non-intersection crash concentrations. For this research, crash information was integrated with roadway feature information to produce tabular listings organized by increasing mile point, providing a rich information environment for safety engineers addressing safety issues on any given highway section. Additional information can be extracted from a photolog database, eliminating to a large degree unnecessary field visits. In addition, overall crash rates, crash densities and other statistics were developed for each number-of-lanes/population density (urban or rural) cohort. Statewide statistics were developed as well. Specific injury and fatal, wet and snow, darkness, horizontal or vertical curve, and fixed object crash statistics were produced. The methodology developed to address statewide non-intersection crashes systematically provides a powerful tool to prioritize statewide safety improvements, apply for federal safety improvement program funding, and address litigation cases.
Abstract