The Design Of Data Analysis System Based On Operational State In Smart Transportation

Whether it is the daily work or entertainment, people’s lives are always very dependent ontransportation. With the growing global population, advancing transportation facilities and vehiclesstill can’t keep up with the growing traffic flows. So traffic problems can be described as aworldwide problem. Intelligent Transportation System encountered a number of bottlenecks in theworldwide application. At the same time, the Smart Transportation System derived from IoT bringsnew hope to solve the traffic problems. But the unfolded Smart Transportation System brings morecomplicated and big traffic data as well. The problem is how to get some information useful fortraffic management from these big data. Based on this hot topic, the thesis proposes a distributedparallel-Data Mining solution to deal with the big traffic data.This thesis combines the literatures from home and abroad to compare different technologiesand analyzes the characteristic of information at every level of smart transportation systems, andgives an intelligent transportation system architecture that contains the data connection layer, thenetwork transport layer, the data analysis layer and the application layer. In the step of data analysis,according to the characteristics of traffic data, this thesis comes up with a new traffic data miningsystem based on a new distributed and parallel technology——Hadoop. And based on the aboveinnovative idea, this theies designed a data analysis system based on operational state in smarttransportation. With the deployed Hadoop cluster, this thesis used MapReduce programming modelmixed with the improved Apriori algorithmto run an example of large traffic data, and digged out afavorable information to traffic control. It proved the feasibility and effectiveness of using Hadoopplatform and data mining to solve big transportation data.This thesis summarizes the points which can be improved in the example and looks into thefuture of intelligent transportation systems based on data mining. It is believed that the distributedparallel data mining solution will certainly contribute to intelligent transportation systems in thefuture.