AN ADVANCED METHOD OF IDENTIFYING HAZARDOUS LOCATIONS

Selecting individual highway facilities for an in-depth safety analysis and potential safety improvement is one of the key functions of the Safety Management System (SMS). Since this process includes all road facilities administered by the SMS center, the selection process must be automated due to a typically large number of candidate locations. The methods currently in use rank locations according to crash rates or number of crashes. They deploy the criteria that detect locations with abnormally high number of crashes. As it is shown in this paper, the current methods may not ensure the maximum crash reduction - the objective of the SMS. This paper proposes a new method that overcomes the weakness of the current methods. In the proposed method, potential safety benefit is used as a criterion to rank locations. The potential crash reduction is the difference between the overall crash rate and the minimum crash rate. The minimum crash rate is the part of the overall crash rate that cannot be eliminated by applying safety countermeasures at individual sites. The paper describes the method of estimating both the overall and minimum crash rates. The crash rate estimates take the form of belief functions to incorporate the authors' uncertain knowledge of the true values. To estimate these values, the procedure accepts data of varying scope and quality. The proposed method promotes selection of sites where the potential for safety improvement is high. The paper presents an example application of the method.