Prediction Method Based on Target Trajectory Pattern Matching
Qun Lai, Jinye Yu, Shaoyan Jiang · 2013
A specific target trajectory prediction algorithm based on mobile pattern matching is proposed, which is called PS-Tree algorithm. In this algorithm, historical data of targets' movements generated in monitored region are used for pattern mining, coordinate series of targets are converted to region trajectory series, and frequent moving modes are figured out by building up PS-tree. The algorithm matches current trajectories with the ones in pattern library to forecast the movements of the target when a target moves into monitored region. The simulation results prove that PS-Tree algorithm has low time and space consumption but high prediction accuracy.