Research And Application Of Heterogeneous Vehicle Data Matching Algorithm Based On Multi Source Data Fusion

With the increasing development of smart city,there is a large number of information in the form of data in various fields,data quantity,scale,dimension becomes more and more complex,a variety of related data from different sources are combined with more implicit information mining is an important and complicated and time-consuming work.Data mining technology is a hot research topic in recent years.It is mainly used to extract massive data from databases,extract business related data and process them accordingly,and then get key data information for business decision-making.But when the required information is needed when mining multi-source heterogeneous data,due to the differences between data sources,different data structure,magnitude,dimensions are different,a plurality of data sources directly into the same data set is very difficult to achieve,so the multi-source data fusion is a very important part of.This paper studied the heterogeneous vehicle data matching algorithm,the set data represent a smart city,namely the use of Open Street Map to provide the road data,the Microsoft Corp of Beijing city taxi GPS data,OBD data provided by the Beijing network,the number of fuel consumption data through the sign according to the obtained POI data points of interest and the vehicle body CAN bus acquisition the data of heterogeneous vehicle data matching algorithm in different purposes.The specific work of the paper is as follows:1.Study on the matching method of GPS data and network data,the definition of the map matching algorithm and classification,introduces the current commonly used map matching algorithm,hidden Markov model,map matching algorithm and algorithm based on ST-Matching map,expounds the above two kinds of algorithm principle,and the advantages and disadvantages of this paper focused on the use of data.2.Aiming at the low sampling data selected in this paper,based on the ST-Matching map matching algorithm,we simplify and improve it,rebuild the basic road network,divide the concept of road network as the innovation point,and redefine the algorithm flow of data matching.The GPS position in the segmented path network to obtain the edge candidate set,through the vertical distance between H and the angle obtained six candidate most likely edge,topology construction,access to the best matching result by calculating the transmission probability,and two times the screening and verification of the results.3.This paper uses multi-source heterogeneous data matching algorithm,the fuel consumption of OBD data are also mapped to the network,through the driving track and each section of road network in the taxi driver's driving style,driving speed,driving fuel consumption information,the study of the relationship between fuel consumption and city road.It is foundthat there are some special sections.Because of the reasons of road attributes,such as the length of the roads and the number of traffic lights,the drivers will have similar fuel consumption when they pass the special section.4.This paper uses multi-source heterogeneous data matching algorithm,the interest points POI information is mapped to the network,clustering to obtain new information on the traffic grid through the POI network data,calculate the interest point and linear distance clustering of traffic grid points of interest area,get the passenger hot region,through the Bayesian theorem that city residents travel regularly,and verify the results of.5.This paper uses multi-source heterogeneous data matching algorithm,the vehicle body CAN bus data acquisition is matched to the network,by matching the results of vehicle driving conditions,and the driving conditions in different working conditions and driving the vehicle to determine ACC automatic cruise driving performance difference,so as to improve the system control strategy.Finally,this paper summarizes the work of the paper,summarizes the innovation and achievements in the research process,points out the shortcomings and limitations,and puts forward some suggestions for future research directions.