Title: Research on primary traffic congestion point identification method based on fuzzy logic

Authors: Haitao Guo; Lunhui Xu

Addresses: School of Mechanical and Electrical Engineering, Guangdong Construction Vocational Technology Institute, Guangzhou, Guangdong, China; School of Civil Engineering and Transportation, South China University of Technology, Guangzhou, Guangdong, China ' School of Civil Engineering and Transportation, South China University of Technology, Guangzhou, Guangdong, China

Abstract: This study proposes a method to discriminate primary traffic congestion points based on interval type two fuzzy logic combined with improved generative adversarial network, and proposes quantitative congestion point change criteria by using the difference in spatio-temporal order between primary and secondary congestion, and classifies congestion points into four types: primary congestion, primary dissipation, and secondary congestion and secondary dissipation. The methods are also subjected to comparative analysis and ablation experiments to determine the improvement of the optimisation on performance and efficiency. The experimental results show that the method proposed in the study improves the accuracy by 20% to 89% after training, and the average error rate is only 5.5%, which is better than the mainstream congestion point discrimination methods in terms of convergence and efficiency. Finally, the discriminative law of primary traffic congestion points is summarised with the congestion discriminative results of a city for a week.

Keywords: type two fuzzy logic; generative adversarial networks; dual attention mechanism; primary congestion; secondary congestion.

DOI: 10.1504/IJVICS.2023.131603

International Journal of Vehicle Information and Communication Systems, 2023 Vol.8 No.1/2, pp.135 - 151

Received: 26 Jul 2022
Received in revised form: 12 Dec 2022
Accepted: 25 Apr 2023
Published online: 20 Jun 2023 *