Time Mode Based Next Position Prediction System

Position prediction of moving object has become a reality utilizing the vast amount of location data acquired by positioning devices embedded in mobile phones and cars. In this paper, we proposed a position prediction system which focuses on the time regularity of object moving. Historical location data of the object is used to extract personal trajectory patterns to obtain candidate next positions. Each of the candidate positions is scored by the proposed Time Mode-based Prediction (TMP) algorithm according to the proximity between the time component of patterns and current time. The position with the highest score is regarded as predicted next position. Furthermore, a hybrid B/S and C/S architecture is employed to perform the real-time prediction and results display. An evaluation based on a public trajectory data set of 12 objects demonstrates that the proposed TMP algorithm can realize position prediction with high accuracy. Moreover, the average accuracy rate of our prediction algorithm is about 85.5%, which is 33.7% greater than the Markov-based algorithm with one known position.