Compression algorithm of ship trajectory based on online directed acyclic graph

To solve the problem of ship trajectory data compression, an online ship trajectory compression algorithm was proposed. An estimation method of multiple sliding reckoning the ship position was used to clean the ship trajectory, and the online directed acyclic graph was used to build the compression path tree on the clean trajectory and to output the sampling points. The Hash table was used to manage the trajectory queue and path tree in memory in order to improve their query speeds. To verify the effect of the proposed algorithm, the compression times and errors of real ship automatic identification system, direction-preserving algorithm and Douglas-Peucker algorithm were compared, and the original trajectory, clean trajectory and compression trajectory were analyzed by the visual method. Experiment result shows that in terms of the compression time, the compression times of direction-preserving algorithm and Douglas-Puck algorithm are about 1.1 and 1.3 times of the proposed algorithm, respectively, indicating that the execution time of the proposed algorithm is shorter than the other two algorithms. The proposed algorithm retains the time information in the compression process, and the average synchronized Euclidean distance error is less than 10 m at any compression rate. The maximum synchronized Euclidean distance error is only 127 m when the compression ratio is 1%, while the average synchronized Euclidean distance errors and the maximum synchronized Euclidean distance errors of the other two algorithms can not be controlled, and change randomly. In terms of perpendicular distance error, both the proposed algorithm and Douglas-Puck algorithm can guarantee that the perpendicular distance errors less than 20 m when the compression rate is not less than 5%, while the perpendicular distance error of direction-preserving algorithm changes randomly. In terms of display effect, the proposed algorithm can effectively remove the noise points of trajectory, and the compression trajectory can better represent the macroscopic traffic flow of original trajectory. Therefore, the proposed algorithm can preserve the shape and time informations of original trajectory more efficiently. © 2020, Editorial Department of Journal of Traffic and Transportation Engineering. All right reserved.