Position prediction system based on spatio-temporal regularity of object mobility

• Published in: Information systems (Oxford) 2018-06, Vol.75, p.43-55
• Main Authors: Li, Xin, Yu, Chongsheng, Ju, Lei, Qin, Jian, Zhang, Yu, Dou, Lei, Yuqing, Sun
• Format: Article
• Language: English
• Subjects: Algorithms; Cellular telephones; Data acquisition; Data acquisition systems; Data mining; Electronic devices; Information systems; Location based services; Markov analysis; Markov processes; Mobility; Position (location); Positioning devices (machinery); Predictions; Public transportation; Regularity; Trajectory analysis
• ISSN: 0306-4379; 1873-6076
• DOI: 10.1016/j.is.2018.02.004

•An effective position prediction algorithm for moving objects is proposed.•The algorithm focuses on both spatial and temporal dimensions of trajectory data.•Three data filters are designed to remove redundant data and outliers.•A hybrid C/S and B/S framework is introduced to implement the prediction system. With the accumulation of the vast amount of location data acquired by positioning devices embedded in mobile phones and cars, position prediction of moving objects has been an important research direction for many location-based services such as public transit forecasting and tourist behavior analysis. In this paper, a position prediction system has been proposed, which utilizes not only spatial but also temporal regularity of object mobility. Historical trajectory data of the object is used to extract personal trajectory patterns to obtain candidate next positions. Each of the candidate next positions is scored by the proposed Spatio-Temporal Regularity-based Prediction (STRP) algorithm according to time components of patterns and current time. The position with the highest score is considered as the 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 two different public trajectory data sets demonstrates that the proposed STRP algorithm achieves highly accurate position prediction. Moreover, the average accuracy rate of our prediction algorithm with one known position is about 86.8%, which is 43.9% better than the Markov-based algorithm.