A Fast Identification Model For Expressway Accidental Congestion Based On ETC Data

Accidents on expressways are often more severe than those in urban areas,causing traffic congestion and increasing the risk of secondary accidents.Therefore,rapid detecting accidental congestion on expressways can help traffic management departments respond promptly and alleviate congestion.This is of great significance for improving public travel efficiency,ensuring the safety of drivers and passengers,and avoiding secondary accidents.However,accidental congestion on expressways is characterized by suddenness and randomness and has a high degree of similarity with the traffic flow characteristics of non-accidental congestion,resulting in serious challenges such as high false alarm rates and low detection efficiency in detecting accidental congestion.To address this,this thesis focuses on the ETC system and ETC transaction data in a certain province,and conducts in-depth research on the rapid identification of accidental congestion on expressways.Firstly,this thesis pre-processes the ETC transaction data to ensure data reliability.Then,a congestion identification model is constructed to identify congested sections on expressways using historical data.Based on this,the traffic flow characteristics of accidental congestion on expressways are analyzed,and a fast identification algorithm for accidental congestion on expressways based on ETC data is constructed.The main contributions of this thesis are as follows:(1)This thesis proposes an expressway congestion identification model based on ETC transaction data.By analyzing historical ETC transaction data,a congestion feature model for expressways is constructed,and an improved fuzzy clustering algorithm is introduced to identify the traffic flow status on expressways.Through experimental comparison,the effectiveness of the proposed method is verified.(2)This thesis analyzes the characteristics of accidental congestion from the perspective of ETC transaction data.Firstly,the definition of accidental congestion is given.Then,the traffic flow evolution characteristics of accidental congestion are studied by analyzing ETC transaction data and the traffic flow characteristics of expressways.Based on this,a method is proposed to distinguish accidental and nonaccidental congestion by comparing the speed distribution during both types of congestion and the "first in,last out" phenomenon of accident vehicles.This provides a theoretical basis for the rapid identification of accidental congestion on expressways.(3)This thesis proposes a fast identification model for accidental congestion on expressways based on ETC transaction data.Firstly,the time aggregation method for ETC transaction data is studied.Then,based on the impact of accidental congestion on the traffic flow changes of upstream and downstream sections and the analysis of the characteristics of accidental congestion in Chapter 4,relevant evaluation features are studied,and a feature model for accidental congestion is constructed.Through different experimental algorithms comparison,a fast identification model for accidental congestion on expressways is constructed.The method was verified by real-world data,and the detection efficiency was 98.73%,the accuracy was 98.7%,the detection rate was 95.8%,the false alarm rate was 0.16%,and the average recognition time was 5.2 minutes.This model provides a feasible solution for the fast identification of accidental congestion on expressways.