Citation

Wu CG. Adv. Transp. Stud. 2022; (SI 4): 145-156.

Copyright

(Copyright © 2022, Arcane Publishers)

DOI

unavailable

PMID

unavailable

Abstract

In order to overcome the low recognition rate of road traffic status and low evaluation accuracy in road traffic safety assessment in different time periods, a road traffic safety assessment method in different time periods based on deep neural network was proposed. Firstly, build a road traffic safety assessment index system at different time periods. Then, combined with the bat algorithm, the support vector machine is improved to identify the road traffic status in different time periods. Finally, a deep neural network is used to construct a probability function with the identified traffic state as input, and output the safety assessment result, so as to realize the road traffic safety assessment result in different time periods. The experimental results show that the minimum recognition rate of road traffic state in different periods is 96.31%, the average recognition time is 0.59s, and the minimum evaluation accuracy rate is 96%. It has the advantages of rapid and accurate recognition of road traffic state in different periods and high evaluation accuracy.

Language: en

Keywords

Road Safety; Traffic; Transportation