Research On Moving Objects Trajectories Analyze Technology

With the rapid development of Global Positioning System (GPS) and wireless communication network, traditional location-based service towards moving objects can not meet the needs of users and face many challenges. As one of the most important analyzing techniques of moving objects, trajectory analyzing techniques play a very important role in moving objects location-based services and have broad application foreground. Currently the researches on moving objects'trajectory analyzing are preliminary and far from maturation both in theory and practice. And there are many open problem to solve.In order to study the difficult and hot issues in moving objects databases, this paper gives comprehensive discusses and analysis on former work in related areas. And towards the application needs, this paper studies the key techniques of moving objects trajectories analyze, and proposes the approaches of trajectories pretreatment, ROI (Regions of Interest) discovery and FIP (Frequent Interesting Pattern) discovery. The main work and innovations are detailed as follows:1. This paper proposes PBGP (Prune based on Grid Partition) approach to reduce the amount of trajectory data. According PBGP approach, we mapping trajectories into grid cells, and combine with ROI discovery qualification, we prune cells by the number of trajectories in each cell. In order to guarantee the integrality of result, the approach specially analyzes the boundary area.2. According to investgating the ROI discovery approach, this paper proposes two stage approache to discovery mass moving objects trajectories'ROI. First, according to improved DBSCAN to discovery single trajectory'cluster, and use average location point to donate clusters. Then this approache adopts traditional DBSCAN to discovery mass trajectories'ROI. In improved DBSCAN method, we propose the notion of Eps liner adjacent area, and propose quantile function to decide the parameters.3. It proposes moving objects FIP discovery approache. Based on discovery of ROI, this paper researches the discovery of FIP of single and mass moving objects trajectories. To single trajectory FIP discovery, we build the ROI-list according to the record information of each ROI. And considering the appearance frequency and time of each moving object, we can find the FIP of each moving object. To mass moving objects trajectories FIP, this paper proposes a calculation method based on the interesting degree of ROI and experience degree of moving object. This method can discovery FIP of any length.Finally,according to instance analysis, this paper proves the advanges of the approaches proposed above on time consuming and accuracy.