01657533

Component

In recent years, sliding window method have been widely adopted in western countries to identify hazardous road segments due to its simpleness and practicability. However, if the segment length of sliding window or increment is inappropriate, two common problems may emerge: hazardous road segments are omitted or exaggerated. This paper aims to investigate the optimal lengths of sliding window and increment. The influence of lengths of sliding window and increment on the identification results of hazardous road segments is discussed, and the optimal window lengths and increment lengths are determined by using quantile method and graphic method. Two methods are tested to prove their reliability in a novel and more practical way. Results indicated that: (1) different lengths of sliding window should correspond to different identification criteria; (2) the longer sliding window length is, the greater possibility of exaggerating hazardous road segments will have; (3) the bigger increment length is, the greater possibility of omitting hazardous road segments will have; (4) in respect of determining the optimal length of sliding window and increment by using quantile method, the optimal combination is that increment length is half of the window length. Findings of this paper showed that both quantile method and graphic method are feasible to optimize the traditional sliding window method and when the window length is smaller, the graphic method is more reliable than the quantile method.

This paper was sponsored by TRB committee ANB20 Standing Committee on Safety Data, Analysis and Evaluation.

18-01974

English

17p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

Figures; References (40) ; Tables

Crash causes; Geometric segments; Highway design; Highway safety; Mathematical models; Risk assessment

Highways; Safety and Human Factors

Transportation Research Board Annual Meeting 2018 Paper #18-01974

TRIS, TRB, ATRI

Jan 8 2018 10:29AM