IDENTIFICATION OF ACCIDENT BLACK SPOTS USING NETWORK SCREENING: THE CASE OF SOGUTLUCESME-15 TEMMUZ SEHITLER BRIDGE CORRIDOR

Abstract

In order to reduce the number and the effects of traffic accidents on a roadway, various countermeasures are taken into consideration. The first step to decide the proper countermeasures is identifying Accident Black Spots (ABS) and then improving the site regarding the different type of the countermeasures to reduce the effect of the traffic accidents. As well as many methods used to identify the ABS in literature, network screening technics, that simple ranking, sliding window and peak searching, are defined in Highway Safety Manual published by AASHTO in 2010. In these technics, there are various performance measures like average crash frequency, equivalent property damage only, etc. to rank the roadway segment. Based on the ranking, ABS are identified and prioritize to decide and implement the countermeasures. In this study, data for fatal-injured traffic accidents that occurred at Sogutlucesme-15 Temmuz Sehitler Bridge corridor in Istanbul between 2011-2013 are provided by Istanbul Directorate of Security and the data was transferred into geographical information systems (GIS) with that way the data was related with the geographical location. The corridor was split into 10 m long segments in GIS. Three different types of performance measures are considered to rank the segments based on the above-mentioned network screening technics. K-means clustering method was used to identify the ABS in this study. As a result of the study, the K-means clustering method is accomplished to identify the ABS and sliding window technic is the most appropriate methods to identify the ABS.