Short-term Traffic Flow Prediction Based On Floating Car Data

As the urban population continues to increase,the number of motor vehicles grows too fast,road construction cannot meet the transportation needs in various aspects of structural functions,and the traffic congestion problem is intensifying in China.Therefore,alleviating urban traffic congestion has become an urgent problem in China.The problem.Real-time and accurate prediction of traffic flow is a necessary condition for the development of intelligent traffic management system.It can provide real-time and effective information for the traffic system,enabling the system to achieve path guidance,alleviate road congestion and reduce the emission of harmful gases from motor vehicles.Therefore,the research on short-term traffic flow prediction technology has far-reaching theoretical and practical significance.Most of the existing traffic flow prediction models can't fully study the essential characteristics of traffic flow data,and because of the characteristics of the transportation system,the data volume is large and the feature dimension is high.In order to realize real-time and accurate short-term traffic flow prediction,and in view of the nonlinear time-space correlation expressed by short-term traffic flow data,based on the traditional deep belief network,this paper adds Gaussian noise,considers the prediction task contribution of different nodes,and stacks the new CRBM.The regression is connected at the top level,and the FR-CG algorithm is used to perform global fine-tuning to form a CDSHybrid model to predict short-term traffic flow.Through the screening of the taxi traffic data of Beijing Second Ring Road,the data of abnormal data repair and noise removal,two traffic flow parameters of average speed and traffic volume are extracted.The short-term traffic flow prediction model designed in this paper is used respectively.Forecast traffic volume and average speed.In this paper,a fuzzy hierarchical traffic congestion recognition model based on information entropy weighting is proposed to realize traffic congestion prediction based on traffic flow prediction.Finally,the experimental verification shows that the traffic flow prediction model proposed in this paper is superior to the commonly used traffic flow prediction models in terms of traffic volume and average speed.The traffic congestion prediction model proposed in this paper is validated,and the model has advantages in accuracy and real-time.