Title: Research on feature extraction of vehicle abnormal driving behaviour based on 5G internet of vehicles

Authors: Wei Yu

Addresses: School of Transportation Engineering, Changsha University of Science and Technology, Changsha, 410004, China

Abstract: In order to overcome the problems of high response time delay and low accuracy of traditional vehicle abnormal driving behaviour feature extraction methods, the paper proposes a 5G car networking-based vehicle abnormal driving behaviour feature extraction method. This method uses the infrastructure and sensor equipment of the perception execution layer of the 5G internet of vehicles to collect vehicle information, and selects the Internet Protocol Version 6 communication protocol of the network transmission and control layer for data transmission. Based on the data received by the integrated application layer, the Riemannian manifold method is used to extract the characteristics of abnormal driving behaviours such as emergency acceleration, emergency deceleration, speeding, and frequent lane changes. Experimental results show that the extraction accuracy of this method is as high as 98%, which can effectively extract the characteristics of abnormal driving behaviour, and the response delay during data transmission is less than 80 ms.

Keywords: 5G; internet of vehicles; cars; abnormal; driving behaviour; feature extraction.

DOI: 10.1504/IJVD.2021.122256

International Journal of Vehicle Design, 2021 Vol.86 No.1/2/3/4, pp.124 - 142

Received: 31 Jul 2020
Accepted: 12 Mar 2021
Published online: 14 Apr 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article