Research On Tourism Vehicle Route Planning Technology Based On Big Traffic Data

At present,big data is used more and more widely in the field of transportation.For example,it can accurately reflect the changes of people,vehicles and roads as time and space by synthesising traffic big data of different regions,different areas and different departments;traffic big data helps to better configure traffic resources,achieve traffic management dynamically and improve the utilization rate of traffic resources on road;it can improve the accuracy of identifiying high-risk road in order to help vehicles to avoid obstacles in time to ensure traffic safety by analyzing the traffic big data such as motor vehicles'speed,drivers'behaviors and so on,.In the current researches and applications of traffic big data,there are the following problems.First,traffic big data is not used for a travel plan among multiple cities.Second,when using traffic big data to predict the traffic speed on highways,little sample data will reduce the prediction accuracy.Third,using the data such as motor vehicles' speeds,drivers' behaviors cannot accurately identify areas of obstacles on the highway.Aiming at the problems above,this thesis proposes three algorithms including travel route planning,traffic speed predicting on highway,and obstacle identifying.The specific research contents are as follows:(1)This thesis proposes an algorithm of travel route planning with time optimizing.In order to optimize the driving route of vehicles and reducing the time consumption on highways,this thesis proposes an algorithm called OPLP(Overall Planning inter clusters and Local Planning intra clusters)for travel route planning.The algorithm firstly uses K-Means clustering algorithm to cluster all the travel cities.After that,Hungarian algorithm is used to plan the transfer routes among clusters.Finally,the Greedy Algorithm is used to plan the travel route among cities within one cluster.OPLP can improve the scientificalness and rationality of travel route planning and reduce the time overheads on the highways,and improve the comfort level of travel.(2)This thesis proposes an algorithm to predict traffic speed on highway sections.In order to optimize the driving speed of vehicles on highway sections in order to select the appropriate departure time and the sections for driving,this thesis proposes an algorithm for predicting traffic speed on highway sections.For the traffic speed in holidays,SPA(Segment Prediction Algorithm)is proposed.SPA firstly uses the K-Means clustering algorithm to cluster the data of traffic speeds for the holidays in the previous three years.Then,for the time periods when the speeds change slowly,the value of the traffic speed at each moment is predicted by the weighted means of traffic speeds in the previous three years at that time.For the time periods the speeds change violently,a binary linear model is established according to the linear regression algorithm based on the change trend of speeds within one hour.And then the traffic speed at each moment in this hour is predicted.For the traffic speed in workdays and weekends,because the data volume of traffic speeds on highway sections is large,and there are fixed rules in the data,it has high prediction accuracy by using the LSTM(Long Short-Term Memory).Therefore,this thesis uses LSTM to predict the traffic speed on highway sections on the workdays and weekends.At the same time,in order to eliminate the redundant data,RDRW(Redundant Data Reducing algorithm based on sliding Window)is proposed.RDRW can accurately predict the traffic speed of the highway at every moment,so as to select the appropriate departure time and the highway section to avoid traffic congestion.In addition,RDRW greatly reduces the amount of the redundant data and improves the speeds of model training with little negative influence on prediction accuracy.(3)This thesis proposes an algorithm to identifying the area of an obstacle in a environment that the sensors are deployed sparsely.In order to locate the area of an obstacle in highways under the environment that sensors are deployed sparsely,this thesis proposes an algorithm named MSPT(Minimum Surrounded Polygon identifying algorithm based on visual area Tracking).MSPT uses LED lights to find the minimum convex polygon surrounding the obstacle,so that the area of the obstacle on the highway can be located.When locating the obstacle's area,MSPT firstly looks for the initial positioning light.And then the first edge of the minimum convex polygon surrounding the obstacle is found.Finally,all subsequent edges of the minimum convex polygon are found.MSPT can enable vehicles to locate the areas of obstacles on the highway in time,so as to take measures to avoid the obstacle in time.