A Study Of Trajectory Matching Algorithm Based On Mobile Big Data

With the rapid development of wireless communication and sensor technology,the coverage of communication devices has been increasing,and billions of mobile data can be generated every day.At the same time,the possible range of individual activities has gradually expanded.Due to the diversity and regularity of moving object,the trajectory data mining has research value,including the research of trajectory similarity matching in this paper.We can use the mobile big data to implement the trajectory similarity matching.However,as the huge amount of mobile data and the constraints of matching performance,we should build big data environment,propose new trajectory compression algorithm and similarity algorithm to implement the trajectory matching.The main research content of this paper is as follows:1.As the vast amount of trajectories,this paper proposes a multi-dimension trajectory compression algorithm based on important track points to compress the trajectory.The coarse trajectory compression is based on time-segment aggregate approximation,and preserves the location and rotation angles information of track points.The fine trajectory compression makes the trajectory divided by time segments.With the time threshold and distance threshold,we divide the trajectory segments into sub-segments.Then,the sub-segments are weighted aggregated according to the importance of track points.The experimental results show that both trajectory compression algorithms can represent the shape of trajectory intuitively.Meanwhile,the algorithms reduce the quantity of track points and the better trajectory analysis efficiency can be gained.The algorithms also can preserve the key features form original trajectory,the fine trajectory compression can keep key features at 97.98%,and the coarse trajectory compression keeps key features at 93.59%.Because the coarse trajectory compression holds small qua ntity of trajectory data,and the fine trajectory compression preserves enough key features,they can be used into different research background.2.In trajectory similarity matching,this paper presents a hierarchical similarity matching model based on structure clustering and information entropy interpolation.We choose a target trajectory,and then,build a two-layer similarity matching model.The first layer uses coarse compression trajectories,and then filters out the similar trajectory set through the result of structure clustering algorithm.The second layer,we make the trajectories fine compression expression,which are in the similar trajectory set.Then,the important mobile trajectory segments are selected by information entropy.We calculate the Hausdorff distance based on temporal interpolation between these trajectories and the target trajectory.Finally,we get a number of the most similar trajectories to the target trajectory,and each of them has a similarity score.The experiments show that the algorithm has the merits of high calculation efficiency.At the same time,compared with the Hasudorff distance method and the Euclidean distance method,the algorithm in this paper has higher accuracy,and it can give objective similarity score.3.In order to verify the effectiveness of the algorithms in this paper,we design and implement a trajectory similarity matching system.We implement big data storage structure with Mongo DB sharding.Then,we build the data processing platform to provide data service automatically and regularly,including trajectory extraction,trajectory data cleaning and multi-dimension trajectory compression.The hierarchical similarity matching model is used to trajectory similarity matching.It helps realize the function,including trajectory operation,trajectory similarity matching and other trajectory similarity matching functions.The system can manage the big data safely and efficiently.It can predict and filter out the suspicious groups to improve the social security,and it also can look for the hotspot region and regional correlation to provide daily wireless location service.Now,the system has been running stably,and has good practicability.