SAFETYLIT WEEKLY UPDATE

We compile citations and summaries of about 400 new articles every week.
Email Signup | RSS Feed

HELP: Tutorials | FAQ
CONTACT US: Contact info

Search Results

Journal Article

Citation

Jing HL, Ye LT, Wang JZ, Xie Z, Brown M. Adv. Transp. Stud. 2018; (SI 2): 15-24.

Copyright

(Copyright © 2018, Arcane Publishers)

DOI

unavailable

PMID

unavailable

Abstract

The conventional gray predication model GM (1, 1) cannot accurately analyze the dynamic traffic index information of complex and scattered road sections because it may cause relatively large error and performs not well in stability. In order to solve this problem, a dynamic traffic safety grade evaluation model for road sections based on gray fixed weight clustering is designed. In this method, In this method, the gray clustering evaluation method is adopted for gray clustering to complex and scattered traffic safety grade evaluation indexes, and the gray fixed weight clustering method is adopted to weight each clustering index in advance; the clustering weight of each index is set by a fuzzy consistent matrix, on which the fixed weight coefficient of the index is calculated and the clustering vector is constructed; the cluster coefficients and cluster vectors are combined to obtain the clustering indexes of traffic safety evaluation; then a BP neural network dynamic traffic safety grade evaluation model for road sections is constructed according to the indexes, so as to accurately evaluate the dynamic traffic safety grade of road sections. The experiment results show that the designed model method can effectively evaluate the dynamic traffic safety grade of 31 road sections in areas with a high probability of traffic congestion with small evaluation error and high stability, so it meets the design requirements.

Keywords: gray fixed weighted clustering; complex and scattered; clustering weight; BP neural network; road section; traffic safety grade


Language: en

NEW SEARCH


All SafetyLit records are available for automatic download to Zotero & Mendeley
Print