Speed profile variation as a road network screening tool

Roadway sections that experience a higher-than-expected number of crashes are usually identified by long-term crash frequency data. When historical crash data are either limited or unavailable, surrogate safety measures based on roadway characteristics (e.g., road geometry, traffic volume, and speed) are often substituted. This study developed and evaluated several candidate measures to estimate crash frequency on urban streets on the basis of speed consistency along the roadway. These speed consistency measures were based on speed profiles along road segments collected from vehicles equipped with Global Positioning System (GPS) devices. The relationships between these surrogate measures and historical crash frequency were quantified with the use of a combination of regression tree and generalized linear modeling approaches. The findings support the use of the profile-based measures to evaluate the safety of road networks as the deployment of GPS-equipped vehicles becomes more prevalent.