SAFETY IMPROVEMENTS AT HAZARDOUS SIGNALIZED INTERSECTIONS

Abstract

Signalized intersections are supposed to have no safety issue due to the presence of traffic signals controlling various movements; however, the real fact shows something different than the expectation. In 2018, 13,726 accidents occurred at signal-controlled intersections, killing 3,325 persons and causing 12,468 injuries. Fundamentally, this research project will help to identify the sites with promise correctly by applying advanced and precise statistical methods. For the current study, 14 signalized intersections are identified from different places of Delhi city. Geometric data, traffic operation data, and crash data are the three main datasets developed and analyzed in aggregated and disaggregated manner. The percentage share of each crash type, severity, and time of occurrence to the total number of crashes was identified. It was observed that pedestrian and rear-end collisions have more than 50% share and the highest proportion compared to other crash types. Daytime crashes show a 57% share of the total number of casualties. The potential for Safety Improvement (PSI) method was employed, and riskier intersections were identified. PSI is the difference between the annual expected number of crashes and the predicted number of accidents for each selected location. Advanced statistical methods were applied instead of the traditional approaches to overcome the Regression-to-the-Mean (RTM) bias and the effect of traffic volume variation. Empirical Bayes (EB) approach and safety performance functions (SPF) were used to estimate the number of expected and predicted accidents, respectively. The Positive PSI values imply that the site(s) are unsafe; in contrast, the zero or negative values indicate that the location is comparatively safe and there no urgent need for any improvements. Eleven junctions were identified as unsafe and needed to improve. Several crash causation factors leading to the accident occurrences were identified and thoroughly studied. Several countermeasures were proposed to reduce the frequency and severity of accidents at each location based on the influencing factors. A benefit-cost analysis (BCA) was used to evaluate each suggested countermeasure economically. This study will be helpful in addressing the safety issues at signal controlled intersections and assist in the effective utilization of public funds.