Research On Contour Maintaining Oriented Trajectory Compression Algorithm

With the popularity of intelligent mobile terminals and a variety of positioning-enabled devices,people can easily get the individual's mobile trajectories.These trajectories are the basis of many location-based services,which contain a large number of valuable knowledge patterns.Through the analysis of the trajectory data,it can provide support for user profile analysis,hot-spots region detection,urban traffic monitoring and urban functional area identification and etc.Usually the raw trajectories contain a large number of redundant data,the size of raw trajectory data is very large.And with the increase in the number of users,the amount of trajectory data is almost exponentially increasing.Massive volumes of trajectory data will not only take up a lot of storage space,but also bring great challenges to data transmission and data processing.Trajectory compression technology can significantly reduce the space required for trajectory data storage under the premise of losing a small amount of information,and facilitate the data transmission and processing.Traditional position-preserving compression algorithms focus on capture trajectory's position information,and it liable to result in the loss of moving object's direction information in the compression process.However,the direction information of the trajectory is important in describing the semantic meanings of trajectory,the direction change of the moving object indicates the behavior of the user(staying.photographing,etc.).And many trajectory-based discovery tasks rely heavily on the direction information of the trajectory.such as trajectory clustering and classification,map matching and outlier detection.In trajectory compression algorithms,online compression algorithms has the advantage of support online compression,so it plays an important role in trajectory compression algorithms.But for the existing online compression algorithm,it not only has the problem of direction error uncontrollable,it also has the problem of a larger Synchronous Euclidean Distance(SED)error and a lower compression accuracy.In order to further reduce the SED error of the online compression algorithm.in this paper,an improved algorithm Local Optimum Open Window Time Ratio(LO-OPW-TR)based on Open Window Time Ratio(OPW-TR)algorithm is proposed.By using the new anchor selection strategy based on local optimization,the online algorithm LO-OPW-TR can effectively reduce the SED error.In order to solve the existing online algorithm's problems of loss of moving object's direction information,loss of inflection points and poor contours preserving.The LO-OPW-TR algorithm has been further improved,a heading maintaining oriented trajectory compression(HMOTC)algorithm is proposed in this paper.The HMOTC algorithm not only taking into account the heading change degree of positioning points,but also taking into account the direction change degree between positioning points and trajectory segment.So,this algorithm can accurately capture the direction information of the trajectory.By adding the control ability of the direction error,which makes the algorithm can achieve a better control of the trajectory contour and a more accurate approximation of trajectories than traditional position-preserving trajectory compression algorithms.Experimental evaluation with real-world dataset and a variety of error metrics shows the benefit of the HMOTC algorithm.