Research On Mining The Trajectory Patterns Of The Moving Objects For GPS

Due to the fast development of economy in China, the traffic congestion has become a common phenomenon in a metropolis. Therefore, finding a convenient route to the destination in a short time has become people’s urgent needs. In general, people tend to choose the path which is known to the public according to the navigation in path planning. However, in dense populated urban area, because of the differences among the residents around as well as the living focus of various places such as schools, kindergartens, markets or else, the paths recommended by the navigation are not always efficient. People who live in a relatively fixed region are more familiar with the regulation of the living focus around and the road condition at different periods. As a result, they regularly tend to choose paths where the road condition is better, thus avoiding the possible periodic traffic jam near markets or schools. Concerning this issue, the paper launches a research by analyzing the trajectory sequence pattern of the moving objects in order to find out roads that are regularly in good traffic conditions. The purpose is to make reasonable recommendation for users to choose paths to go. The task mainly includes the following aspects.In order to solve the above practical problems, we have made the following study by using large datasets.First, in terms of trajectory data preprocessing, this paper presents a grid-based data processing method for the problem of large amount of data, high time complexity and other characteristics. The time period is divided into different time periods to establish a hierarchical method using adaptive lattice structure and the time period of the data is mapped to the grid. Then a user identification is created based on different track data of inverted index lists with different grid represent a track identification for processing the next track.Secondly, in terms of extracting trajectory feature for each of the high-hierarchy grid cell on a different time segments, we extracted time and location information of the different high-hierarchy grid cells. And then we extracted the effective features as the standard of judging traffic congestion for the grid, which then prepares for mining the trajectory sequence pattern.Thirdly, in the aspect of mining periodic sequence patterns and connecting connected sequence pattern, this research presents trajectory analysis based on individual user location information in different time periods. At first, this research judges whether the trajectory sequence patterns could meet the standard of traffic regularity, then determines whether the extracted sequence patterns are valuable periodic sequence patterns. According to the head and tail of the periodic sequence patterns, we can determine whether the adjacent grid cell can be connected. If so, it will form connected sequence patterns which have more reference significance.Finally, this paper use real GPS track dataset and extended according to the data collected to track a large number of datasets experimental demonstration and analysis. The experimental results show that the proposed moving objects trajectories inverted more primitive methods can greatly reduce the time spent on t finding the periodic trajectory patterns and have further significance in the application of real life.In summary, this paper presents a meaningful problem for searching periodic trajectory patterns. We create a set of inverted index structures of datasets, determine the traffic in the road-network and mine periodic sequence pattern and connected sequence pattern. Moreover, we solve the problem of regular behavior of the user based on trajectory patterns. In this paper, the proposed method does not require any additional relevant information and just consider the trajectory of original features. This has obvious advantages in terms of operability and usability.