Locate The Community On Traffic Conditions By Mining GPS Speed Time Series

The rapid growths of modern cities have caused a series of inevitable problems,such as traffic jams. These traffic problems are not independent. They are closelyrelated to the patterns of human activity. Current traffic stream analysis and predictionmodels can be improved by understanding how human activities affect the trafficstream. A “Community” is an area where a group of related people with some similarcharacteristic gathered together. Different community has different influences totraffic stream. There are only some qualitative sociology definitions of “community”,which are not enough for further research. Therefore some formal definitions andfeature analysis are needed for further research of “community dynamics”. Theachievements of this paper are shown as following:1. This paper uses Internet of Things data to analysis traffic flow. The GPSdata (Global Positioning System) gathered from vehicles contains information ofwhole traffic system, and it has become a research hotpot. The traffic stream timeseries data is a good representation to the behavior pattern of traffic system, and italso reflects the interaction of different parts of the city. So that has become thefoundation of the formal research of community. This paper focuses on mining andsummarizing communities of different types from the traffic stream time series data.2. This paper focus on finding the impact of community on nearly traffic. Someresearchers have shown that, community activities have significant impact to therush-hours, and are the key factor of traffic jam. Roads that are influenced differentlyby communities have different characteristics.3. This paper also focus on finding the impact of community on whole traffic web.By clustering analysis, we can render a Road spectrum for a city. By analyzing theroad spectrum, it can be found that communities have minor influence upon theoverall condition of the road map. The influence of community to traffic has twocharacteristics: time limits and space limits.By analyzing these two characteristics, we can locate communities of differenttypes such as: commercial area, residential area, school etc. That’s the foundation offurther research on the distributions and interactions of communities. This paper has made some research and summarizing to the above problems. And some future worksare also being discussed.