Research On Highway Traffic State Prediction Based On Traffic Parameter Prediction

Highways are playing an increasingly important role in driving economic development and people’s daily traffic.Although related infrastructure is constantly being improved,the rapid growth in traffic volume is still in short supply,causing normal traffic congestion to extend to highways.This has caused a series of problems such as increased energy consumption,frequent road accidents,and lagging management guidance.Up to now,although there have been a lot of theoretical exploration and practical application results on the research of highway traffic state prediction,it is not perfect,and with the continuous deepening of research in related fields such as traffic flow forecasting,such problems are discussed there will be higher requirements.Therefore,this article has launched a study on the method of highway traffic state prediction based on traffic parameter prediction.First of all,based on the detection data of the Los Angeles I-5 highway loop coil provided in the PEMS system,the advantages and disadvantages and applicability of the current traffic information collection methods are counted and discussed,and the traffic parameters are analyzed and studied after the data is preprocessed.Characteristics and time-space correlation,and the specific reasons for this phenomenon.In addition,clarify the relationship between traffic parameters and traffic conditions.Secondly,in order to fully explore the temporal and spatial characteristics of highway traffic flow data,it is proposed to construct a GCN-Bi LSTM combination prediction model.First,use Graph Convolutional Neural network to capture the spatial characteristics of traffic flow data,and then use Bidirectional Long-term Short-term Memory network to analyze the time characteristics of the data to maximize the potential laws of historical traffic flow,so as to accurately predict the characteristic parameters such as flow,speed and occupancy.Then,the application of Fuzzy Clustering in traffic state discrimination is analyzed and explained.In order to predict the highway traffic state more accurately,a fuzzy Cmeans clustering discrimination model improved based on Genetic Algorithm is proposed.First obtain the initial population through FCM,and then use the advantages of Genetic Algorithm global search to find the best clustering center,and then achieve accurate classification,and achieve the purpose of efficiently predicting the traffic state.Finally,in order to verify the feasibility of the above-mentioned construction model,first input the sample data into the prediction model to obtain the predicted values of traffic parameters such as flow,speed and occupancy.It is found that the predicted values are in good agreement with the observed values,indicating that prediction accuracy of the research model is high;then the output of the prediction model is used as the input of the state prediction model to obtain the traffic state classification results.The experimental results prove that the method can effectively estimate the highway traffic state,the usability of the model is confirmed,and it also more intuitively shows the change process of the traffic state of the highway in the future,which is helpful for the early release of traffic information.There are 42 figures,16 tables and 76 references in this paper.