Citation

Huang DX. Adv. Transp. Stud. 2021; 54(SI 1): 35-44.

Copyright

(Copyright © 2021, Arcane Publishers)

DOI

unavailable

PMID

unavailable

Abstract

In order to overcome the problem that the accuracy of the traditional historical traffic accident data modeling method is not high, this paper proposes a historical traffic accident data modeling method based on state observer. The road traffic accident database is established, and the IOT terminal function module and traffic accident monitoring device are designed and constructed in combination with the Internet of things technology, and the invalid data are filtered by cosine similarity in the database to improve the reliability of data modeling. The statistical feature analysis method is used to fuse and cluster the historical traffic accident data, and the distribution concept set of historical traffic accident data is obtained. Based on the obtained data set, a state observer is designed to optimize the modeling parameters, so as to realize the modeling analysis of historical traffic accident data. The experimental results show that the method can still maintain the high accuracy of the data modeling.

Language: en

Keywords

Analysis; Road Safety; Traffic; Transportation