Short-term Traffic Flow Forecasting Method Based On Spatio-temporal Correlation Analysis

In recent years,due to the rapid increase in the number of urban motor vehicles,urban road networks often appear to be congested.The intelligent transportation system came into being,and its core function is traffic control and induction.and real-time and accurate traffic flow prediction information is the key to its traffic control and induction.Therefore,how to obtain real-time and accurate traffic flow prediction information is one of the research hotspots in the field of intelligent transportation in recent years.For the review and analysis of short-term traffic flow prediction literature,the short-term traffic flow prediction research at this stage focuses on predicting the traffic flow on the road time series.The prediction models mainly include statistical theory models,neural network models,nonlinear theoretical models,etc.Since the characteristics existing in the traffic flow space are not considered,there is a problem that the prediction results are not accurate.In this regard,this paper focuses on the consideration of temporal and spatial correlation prediction methods.Firstly,the characteristics of traffic flow in urban road network are analyzed Correlation is found between the time and space of traffic flow.The use of microwave data as a data source was determined by studying the advantages and disadvantages of each data collection method.However,it is inevitable that the data collected by the microwave will be lost or wrong due to equipment aging,detection system error,transmission line failure,etc.Therefore,a set of data preprocessing process is designed,and the data quality is improved after data preprocessing.Secondly,based on the correlation between traffic flow time and space,an improved genetic algorithm is proposed to optimize the BP neural network to predict the traffic flow time series,multiple stepwise linear regression method is used to predict the spatial series,the least square dynamic weighted fusion algorithm is used to fuse the two results to get the final prediction results.Finally,using the measured data of the microwave in the urban road network to analyze the prediction results of the model.From the error analysis indicators of the prediction results,the predicted result after fusion is more accurate than the prediction result of a single sequence.,the validity and accuracy of the fusion prediction model in traffic flow prediction are verified.