Research On Spatiotemporal Propagation Mode Of Urban Traffic Congestion Based On Path Selection Behavior

The traffic congestion caused by urbanization and motorization has become a common "urban disease" faced by major and medium-sized cities around the world.The traditional research methods for urban traffic congestion propagation mainly rely on the analysis of the correlation between traffic flow parameters,making it difficult to distinguish the impact of traffic flow between roads and the true effects of congestion propagation.This article proposes a research method for the spatiotemporal propagation mode of urban traffic congestion based on route selection mode.Based on existing congestion propagation models and route selection pattern mining methods,supported by large-scale motor vehicle travel trajectory data,this study combines congestion discrimination with route selection pattern mining to study systematic congestion spatiotemporal propagation analysis methods and extract traffic congestion spatiotemporal propagation patterns.The main contents of this paper include:(1)Congestion discriminative model based on road network traffic.Through traffic flow data mining,the frequent and occasional occurrence of urban vehicle congestion is discovered.(2)Analysis of the influencing factors of route selection mode.The route selection mode analysis method based on trajectory clustering combines the sequence to sequence autoencoder model Seq2 Seq Auto Encode with the K-Means clustering algorithm.By studying the characteristics of car driving route selection modes,the influencing factors of car route selection modes are analyzed.(3)Method for extracting spatiotemporal propagation patterns of urban traffic congestion.By weighting and resampling the input data,incorporating the characteristics of motor vehicle route selection patterns into the Word2 Vec model,the degree of traffic interaction between two roads is measured,and urban traffic is analyzed and predicted.The analysis structure of the experimental results shows that urban traffic congestion exists temporal periodicity and spatial distribution agglomeration.The spatio-temporal propagation pattern of urban traffic congestion discovered based on the proposed method can provide more comprehensive and scientific support for urban transportation planning and management.