Research On The Recognition And Prediction Method Of The Operating State Of Ring Expressway

As an important transportation facility around the city,the ring expressway expands the urban road network and facilitates people’s daily travel.However,with the increasing traffic demand,the relevant facilities cannot meet people’s travel needs,resulting in the service level of the ring expressway Significant decline,which has put forward higher requirements on the operation and management of expressways.Therefore,the research on the recognition and prediction methods of the traffic flow state of the expressway around the city has practical significance and theoretical value for ensuring the travel demand and quality of residents,and improving the traffic safety and service level of the expressway around the city.The work content and research results of this paper are as follows:(1)Based on the ETC data of the Suzhou Ring Expressway,this paper analyzes the traffic flow characteristics of the ring expressway,analyzes and selects the parameters related to the recognition of the traffic operating state,and proposes a set of preprocessing methods for the traffic flow data of the ETC section.The temporal and spatial characteristics of the traffic flow of the expressway around the city are analyzed in detail,which provides a reference for the selection of traffic state prediction models.(2)Aiming at the characteristics of multi-dimensionality and collinearity of traffic flow parameters,a fusion model of deep autoencoder and DBSCAN clustering algorithm is proposed.The DBSCAN clustering algorithm is embedded in deep autoencoder,and the two-dimensional abstract vector after encoding is used for clustering.It optimizes the clustering effect of DBSCAN algorithm;experimental verification shows that this method can improve the calculation accuracy and reduce the calculation time at the same time.(3)In order to describe the operation state of the expressway around the city in detail,a method of dividing the traffic state based on the RSQ coefficient as the criterion considering the gradient changes of traffic flow parameters(traffic volume,time occupancy,vehicle speed)is proposed.The state is divided into a stable free flow state,a stable low-speed flow state,a meta-stable synchronous flow state,and a meta-stable congested flow state.(4)A method for predicting traffic flow of expressways around the city based on CNN+LSTM neural network is established.The combined model of CNN+LSTM has both the feature extraction characteristics of CNN and the time series adaptability of LSTM.It is different from the direct use of the traffic flow parameters of the past time to map the future traffic state indicators.The traffic flow parameters are predicted first,and then based on The parameter is used to divide the traffic state.Simulation experiments show that compared with the basic LSTM model,the combined model greatly optimizes the accuracy of the model and reduces the calculation time.(5)Taking the sections of the ETC detectors of No.k26+900 and No.k27+700 of Suzhou Ring Expressway as the research object,the traffic state identification and prediction algorithm proposed in this paper is analyzed and calculated,and the effectiveness of the method in this paper is verified.The DBSCAN clustering algorithm improved based on the autoencoder is used to identify the traffic state,and the result meets the threshold of each state.Finally,the LSTM model is used to predict the traffic parameters,and the traffic flow status is identified through the prediction data.The accuracy rate reaches 92.3%,which meets the accuracy expectations.This thesis,through the research on the identification and prediction of the traffic operation status of the expressway around the city,will help the management department of the expressway around the city to accurately perceive the traffic state,improve the management level,and provide a reference for the decision-making of the congestion handling of the expressway around the city.