Search Content

A number of studies have linked the benefits of roadway lighting to a reduction in crash rates at night for a variety of roadway types including rural intersections. Assessing the effectiveness of intersection lighting has primarily relied upon crash database modeling comparing lighted intersections to unlighted intersections. The current research effort gathered similar metrics for comparison but also measured the amount of lighting within isolated rural intersections. Sixty-three intersection locations were chosen for lighting measurement from six different counties within Minnesota. A vehicle mounted illuminance meter data collection system was used to collect data at each intersection. The data collection system utilized five separate illuminance meters and captured horizontal illuminance while driving through all 63 intersection locations. Following data collection, a series of negative binomial regression models were used to assess the horizontal lighting level in conjunction with the nighttime crash ratio, intersection configuration type, and proximity of an intersection to a curve in the roadway. The first model used data from the lighted an unlighted intersections. The results showed that across all intersections, an increase in the average horizontal illuminance (3.91 lux) by 1-lux (~0.09 fc) reduced nighttime crash rates by 9%. A second model used only lighted intersection data and showed an increase in 1-lux from average (6.41 lux) reduced crashes by 20%. A third and final model used unlighted intersections only. A 1-lux increase from average (0.20 lux) or increasing illuminance to lighted levels (as defined by the modeling), reduced nighttime crash ratios by 94%. Intersection configuration and proximity to curves also affected the nighttime crash ratios. The results of this effort are discussed in terms the impact of horizontal illuminance on crash rates. Additional items discussed include thresholds for minimum and maximum lighting levels and future research investigating and validating these efforts.

Run-off-road (ROR) crashes are a concern for two-lane rural and urban roadways throughout Minnesota due to the frequency by which they contribute to fatal crashes (Minnesota Crash Facts, 2013). Mitigating the severity of the ROR events is an on-going research goal in order to help reduce the number of ROR crashes. Examining countermeasures that may reduce ROR crashes is important to determine the most efficient and effective method of warning. Behavioral responses were examined through the use of an in-vehicle haptic-based lane departure warning system (LDWS) using a driving simulator. The study incorporated systematic variation to both the reliability of the warning and sequence of treatment conditions. An additional analysis examined the presence of behavioral adaptation after repeated exposure to the system. Severity of a ROR event was measured as the total time out of lane (TTL) and maximum lane deviation (MLD). Covariates (e.g. road shape) were examined to determine the influence they may have on the severity of a ROR. The results reveal overall LDWS efficacy. TTL was significantly longer when no system was active compared to when it was active. LDWS led to shorter duration of ROR events. Greater velocity was found to be highly predictive of longer TTL. MLD was also greater for baseline drives compared to treatment drives. No behavioral adaptation or system overreliance was detected, suggesting long term benefits of the LDWS. Drivers who actively engaged in a distraction task were at far greater risk of traveling greater and more dangerous distances out of lane.