Evaluation of Bus Driving Behavior Based on Vehicle Location Data

Objective and quantitative evaluation of the driving behavior of bus drivers is crucial to improving the safety of public transport operations. The traditional evaluation of bus driving behavior mainly relies on the subjective evaluation of drivers and regular inspection by regulatory authorities, w...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2023 8th International Conference on Image, Vision and Computing (ICIVC) S. 825 - 830
Hauptverfasser: Zhang, Linliang, Wu, Hongtao, Cui, Kaichao
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 27.07.2023
Schlagworte:
Back-propagation (BP) neural network
Behavioral sciences
Clustering algorithms
driving behavior evaluation
Fuzzy C-Means (FCM) clustering algorithm
Heuristic algorithms
Neural networks
Safety
Training
Vehicle dynamics
vehicle positioning data
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Objective and quantitative evaluation of the driving behavior of bus drivers is crucial to improving the safety of public transport operations. The traditional evaluation of bus driving behavior mainly relies on the subjective evaluation of drivers and regular inspection by regulatory authorities, which has problems such as inconsistent evaluation standards and incomplete data. In this paper, based on the positioning data of public transport vehicles, the characteristic indexes that quantitatively represent the bus driver's behavior are put forward, the bus driver's behavior is classified by Fuzzy C-Means (FCM) clustering algorithm, and then the evaluation model of bus driving behavior is obtained by training with Back Propagation (BP) neural network. Finally, the effectiveness of the evaluation model is verified by traffic accident data. According to the characteristic parameters of over speeding and rapid speed change behavior and the FCM clustering algorithm, four types of bus driving behavior types are obtained: safe type, relatively safe type, relatively aggressive type and aggressive type; the evaluation model of bus driver behavior is obtained by using BP neural network training, and the driver behavior types of two bus lines in Shanghai were identified; In addition, the results of model validity verification show that the vehicles whose driving behavior is judged as relatively aggressive or aggressive account for nearly 80% of the total accidents. The research results will provide a theoretical basis for the safety evaluation of bus drivers and the active intervention guidance of bad driving behaviors.
DOI:10.1109/ICIVC58118.2023.10270432