Research On Route Planning Based On Traffic Flow Big Data

The travel time and path planning of traffic flow prediction is a hot topic in the research of intelligent transportation system.The basic information of urban road traffic flow is the basis of the operation of intelligent transportation system,and it also provides the fundamental support for the optimization and improvement of intelligent transportation system.By studying the basic theory and technology of traffic flow more carefully,we can provide drivers with real-time updated path and accurate travel time,and then achieve the purpose of easing traffic congestion and real-time traffic guidance.With the development of prediction model research and the arrival of big data era,the theory of traffic flow is moving towards the direction of intelligence and digitalization.It is the current and future development trend to carry out big data research in the theory of traffic flow.At the beginning of this paper,several main characteristic parameters of traffic flow big data and their relationship and collection methods are described in detail.According to the properties of characteristic parameters used in this paper,the traffic information data is obtained by using the road traffic video collection technology,and then the traffic flow big data preprocessing method is used to obtain the high-quality traffic flow big data needed in this paper,so as to prepare for the later journey time The calibration of the measurement model and the comparison of simulation provide the fundamental guarantee.Then,with the goal of more accurate estimation of urban traffic section time,all experimental traffic sections in the city are taken as the research object,the classic Kalman filter travel time prediction model is analyzed,and the Kalman filter prediction model is optimized by adding speed limit mechanism into the road supervision speed limit,and the accuracy of the model is verified by the example calculation.At the same time,aiming at many factors of traffic flow big data,thispaper selects attribute subsets,establishes primary and secondary exponential smoothing prediction models according to the characteristics of subsets,and verifies them with the measured data.Finally,based on the Kalman filter,exponential smoothing and data fusion theory,this paper establishes a data fusion prediction model,calculates the weight of each model using entropy method,and makes a comparative analysis of the prediction results.It is extended to all urban road sections,taking the predicted travel time of the data fusion model corresponding to the time scale of each road section as the weight,and using Dijkstra algorithm for path planning to get the real-time shortest time path,which is verified by an example.The results show that the model has a good applicability and the calculation results have a high accuracy.For the prediction of the time of urban road traffic popular journey,the data fusion prediction method can significantly improve the travel time prediction effect.