A Survey for the Ranking of Trajectory Prediction Algorithms on Ubiquitous Wireless Sensors

Abstract

The number of wireless sensors in use—for example, the global positioning system (GPS) intelligent sensor—has increased in recent years. These intelligent sensors generate a vast amount of spatiotemporal data, which can assist in finding patterns of movements. These movement patterns can be used to predict the future location of moving objects; for example, the movement of an emergency vehicle can be predicted for health care decision-making. Although there is a body of research work regarding motion trajectory prediction, there are no guidelines for choosing algorithms best suited for individual needs in uncertain and complex situations and as per the application domains. In this paper, we surveyed existing trajectory prediction algorithms. These algorithms are further ranked scientifically in terms of accuracy (performance), ease of use, and best fit as per the available datasets. Our results show three top algorithms, namely NextPlace, the Markov model, and the hidden Markov model. This study can be beneficial for multicriteria decision-making for various disciplines, including health care.

Localization with Transfer Learning Based on Fine-Grained Subcarrier Information for Dynamic Indoor Environments

“…Indoor localization systems play increasingly important roles in many emerging applications, such as indoor navigation [ 1 , 2 ], disaster rescue [ 3 ], elderly care [ 4 ], etc. [ 5 ].…”