Research On Travel Time Prediction And Traffic Operation Status Discrimination Based On Travel Plan Data

In recent years,the level of urbanization in China has developed rapidly,the number of residents’ motor vehicles has increased sharply,and the supply of transportation infrastructure has become more and more limited.In addition,the opacity of road traffic conditions makes residents have a certain degree of blindness when traveling.Traffic congestion has seriously affected the sustainable development of cities.Fortunately,the country has gradually implemented big data disclosure policies and autonomous driving technology has matured.This makes it possible to release travel plans and obtain user travel plan data before travel in the future.In this paper,the user’s travel plan data is used as the basis of research.First,this paper optimizes the traditional road resistance function to calculate the travel time of users at different spatial nodes;second,in order to judge the traffic operation status of the road section,a traffic operation status determination model is constructed in this paper;Finally,this paper introduces the user travel plan data into the above two models to predict the travel time required for each road section that the user will travel through in the future and to predict the traffic operation status when the user arrives at each road section in the future.The main results include:(1)In order to calculate the road travel time more accurately,firstly,the classical BPR(Bureau of Public Roads)impedance function model is improved.Secondly,an LSTM(Long Short-Term Memory)recurrent neural network is established to predict the positive and negative values of the pending coefficients β in the improved road impedance function.Finally,an improved BPR road resistance function model based on LSTM neural network is obtained.The computational results of the improved travel time prediction model are compared with the computational results of the classical BPR function model,the results obtained by using only the LSTM neural network prediction,and the data actually collected,and it is found that the improved model has higher accuracy.(2)In order to more accurately discern the traffic operation status of the road,firstly,this paper uses the fuzzy C-mean clustering algorithm to cluster the collected traffic data into four categories of states: smooth,basically smooth,slow,and congested.Secondly,three groups of parameters,flow,speed and occupancy,are used to create the basic cloud model for each of the four categories of traffic operation states.When judging the traffic operation status of the road section,the X-cloud generator is used to solve the affiliation values of these three groups of parameters corresponding to each of the four traffic operation states based on the input flow,speed,and occupancy values.Then,the weighted affiliation values are obtained using the entropy weighting method.Finally,the traffic operation status of the road section at that moment is judged.Comparing the calculation results of the constructed traffic operation state discriminative model with the prediction results of LSTM neural network,it is found that the model constructed in this paper is more accurate in discriminating the traffic operation state.(3)In this paper,we use the user’s travel plan data to construct an algorithm that can predict the travel time required for each road section that the user will pass through in the future and determine the traffic operation status of each road section in the future.Then,the Vissim simulation software is used to simulate the traffic on the test section,and the Vissim simulation data is processed and used as the user’s travel plan data to verify the algorithm.The results show that the algorithm combined with the user’s travel plan data has high applicability and reliability in predicting the travel time and discerning the traffic operation status.This paper verifies the accuracy of the two models after establishing the travel time model and the traffic operation status discrimination model,and then fully combines them with the user’s travel plan data.It enables users to anticipate the future traffic operation condition in advance.The method can assist users in making appropriate travel plans.At the same time,the method can help traffic managers to know the usage of expressway in advance,and traffic managers can make the traffic management plan of expressway in advance,so that the usage of expressway can be further improved in efficiency.