As global positioning system is widely used in onboard navigation system, intelligent transportation system has gradually come true in China. However, as the GPS onboard navigation system will create huge amount of traffic information which will take so much time to process all the information and will affect the effect of traffic forecasting accordingly. So data mining from huge amount of GPS traffic data which will get effective information to support short-term traffic forecasting, serve the public to have a nice travel and control traffic condition for the department in charge, will be a meaningful field.The paper is set in the branch of technology project item . " analysis and reasoning of GPS historical data", and does some research on optimization of fundamental traffic information based on data mining from GPS historical data in the following respects. The paper first overview current studies on traffic flow forecasting, collection of traffic information and processing of traffic information (chapter 2). And then advance the frame and models of this study based on current study (chapter 3). After that the paper testes and verifies effectiveness of the research models by a whole year's GPS traffic data of three major trunk roads in Hangzhou, Zhejiang province(chapter 4), and finally come to the conclusions which paves good ways for future researches(chapter 5). |