Research On Reliability Prediction And Discrimination Method Of Urban Road Traffic Congestion Status

As the number of motor vehicles in the city continues to increase,the social and economic problems caused by urban road traffic congestion are becoming more and more serious.In order to effectively alleviate traffic congestion and find scientific and efficient traffic management methods and strategies,it is necessary to predict the road traffic congestion state and clarify the actual traffic congestion status level.This paper analyzes the traffic congestion characteristics of Yangzhou City,establishes the models for predicating average congestion state and suggesting the predication reliability;Constructs the congestion level discrimination method in accordance with the actual situation of Yangzhou City;Finally,based on the congestion prediction model and the discriminant method,analyzes the traffic situation of the roads with the highest congestion in Yangzhou City.In this paper,the average congestion prediction model is established by Autoregressive Integrated Moving Average Model(ARIMA)and the congestion reliability prediction model is established on Generalized Autoregressive Conditional Heteroskdasticity(GARCH).This paper applies a fuzzy comprehensive evaluation method based on the trapezoidal membership function to discriminate the traffic congestion state level.The main research contents are as follows:1.Investigate the status quo of social,economic and roads in Yangzhou City,and find out the causes of normal congestion and sporadic congestion in Yangzhou City through traffic flow data mining and analysis,and summarize the characteristics of traffic congestion in Yangzhou City.2.A time series-based prediction and reliability predictive analysis model for urban road congestion delay index in short-term is constructed,and the performance of the prediction results is evaluated.3.Establish the evaluation index system of traffic congestion status in Yangzhou City,apply the fuzzy comprehensive evaluation method based orn trapezoidal membership function,and discriminate the traffic congestion level based on the actual data and prediction results of typical road sections and regions.4.The typical influencing factors of Yangzhou City traffic congestion are selected for example verification,and the predicted congestion state level is compared with the actual situation to verify the reliability of the model.The study results show that:1.The normal congestion of roads in Yangzhou City is mainly related to factors such as imperfect road infrastructure,unreasonable road network layout,and unscientific allocation of traffic facilities.The occasional congestion is mainly due to weather,construction,holidays,traffic accidents,traffic violations and any other factors.2.The prediction performance evaluation results show that the precision of the ARIMA-based mean prediction model is generally over 90%,the invalid coverage of the GARCH-based reliability prediction model is within 6%,and the confidence interval width is about 0.5,which has a high reliability prediction.3.Through the empirical analysis of traffic congestion state prediction and grade discrimination,it is concluded that the road traffic around Yangzhou scenic spot presents cyclical congestion during different holiday periods;Different climatic conditions have a significant impact on road traffic congestion;The traffic congestion state prediction results for typical road sections are basically consistent with the actual situation,which the trend of change is basically the same.