Title: Driving style recognition of highway-driving semi-trailer at different altitudes

Authors: Ke Liang; Huasheng Chen; Yu Ye; Song Zhang; Mingzhang Pan

Addresses: College of Mechanical Engineering, Guangxi University, Nanning, Guangxi, China; Guangxi Key Laboratory of Manufacturing System & Advanced Manufacturing Technology, College of Mechanical Engineering, Guangxi University, Nanning, Guangxi, China ' College of Mechanical Engineering, Guangxi University, Nanning, Guangxi, China ' Guangxi Yuchai Machinery Group Co., Ltd., Yulin, Guangxi, China ' Guangxi Yuchai Machinery Group Co., Ltd., Yulin, Guangxi, China ' College of Mechanical Engineering, Guangxi University, Nanning, Guangxi, China

Abstract: Driving style provides information about driving behaviour and the driving environment, which reflects the driver's operation while driving. High altitudes can significantly influence the human body, thereby affecting driving ability. Consequently, accurately recognising driving styles at different altitudes has significant implications for driving safety, road design and fuel economy. This paper proposes a method that incorporates data processing, feature selection, a Bi-LSTM autoencoder and spectral clustering to address this issue. Based on the analysis of real-driving data experiments, three driving styles were identified as calm, moderate and aggressive. These styles accounted for 46%, 19% and 36% in plateau driving and 33%, 29% and 38% in plain driving. The results demonstrate how the proposed method can effectively recognise driving styles at different altitudes with fewer features. Additionally, driving styles remained relatively consistent for the same driver driving at varying altitudes, despite changes in vehicle performance.

Keywords: driving style recognition; whale optimisation algorithm; feature selection; Bi-LSTM; autoencoder; spectral clustering.

DOI: 10.1504/IJVICS.2024.136267

International Journal of Vehicle Information and Communication Systems, 2024 Vol.9 No.1, pp.34 - 59

Received: 02 Dec 2022
Accepted: 20 Apr 2023
Published online: 25 Jan 2024 *

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