Short-Term Traffic Flow Prediction And Application Based On Long Short-Term Memory Neural Network

With the development of technology,smart phones are becoming more and more popular,and people's travel habits are changing gradually.In the past,people relied on maps and road signs to get around,but now they use their smart phones to navigate to their destinations.Timely and accurate short-term traffic flow prediction can provide reliable theoretical support for navigation path planning,avoid navigation paths that may cause congestion,reduce driving time,and achieve the purpose of energy conservation and emission reduction.In addition,the short-term traffic flow prediction can provide a reliable theoretical basis for the decision-making of intelligent traffic system.For example,shortterm traffic flow prediction can play an important role in traffic light timing and traffic guidance and so on.Hence,how to get the accurate traffic flow prediction result in a limited time has become the focus of current research.This paper focuses on short-term traffic flow prediction based on Long Short-Term Memory(LSTM).Firstly,the short-term traffic flow prediction based on LSTM neural network model is studied.Secondly,the short-term traffic flow prediction with spatial correlation is studied.Then,in order to further improve the accuracy of traffic flow prediction,this paper studies the short-time traffic flow prediction based on multiple model fusion.Finally,a traffic flow prediction simulation platform is developed based on Android system.We inprove the prediction model,and the data collected from the roads in the Nan'an District is used to verify the reliability of the algorithm.The main work of this study includes these following aspects:1.Short-term traffic flow prediction based on LSTM neural networksThe memory unit of LSTM neural network can store data features in a period of time,which makes LSTM neural network suitable for time series process.Therefore,LSTM neural network is used to predict traffic flow parameters in this study.2.Considering the spatial correlation of traffic flow,the spatial correlation of traffic flow is used in traffic flow prediction processThe Spearman rank correlation coefficient is used to quantify the spatial correlation of adjacent segments in this study by using traffic flow prediction model based on LSTM,and the weights of adjacent segments are obtained by spatial correlation.Furthermore,the influence of adjacent single parameter traffic flow time series on predicted traffic flow is applied to the prediction process,and the model proposed in this section is compared with other short-term traffic flow prediction models.3.In order to further improve the accuracy of traffic flow prediction,the short-time traffic flow prediction based on multi-model fusion is studiedOn basis of Section 1 and Section 2,this paper proposes a short-term traffic flow prediction model based on multiple model fusion,which can fuse the prediction results of prediction models proposed in chapter 2 and 3 to improve the prediction accuracy.4.Development of traffic flow prediction simulation platform display on Android systemIn order to verify the validity of the proposed algorithm,the roads of Nan'an District are selected for study.The traffic flow prediction simulation platform is developed based on multiple model fusion algorithm,and we use the collected traffic flow parameters to verified predicted performance of the simulation platform.