Recently, societal attention on elderly health care has significantly increased as global aging has accelerated. Much effort has been made to apply Information and Communication Technology (ICT) in the provision of elderly health care services. Elderly tracking, or monitoring the presence information of an elderly person, is one such service. This thesis aims to propose an energy-efficient elderly tracking algorithm. It provides a review of some existing methods on positioning as well as some previous works on people tracking and activity recognition. Our proposed tracking algorithm significantly reduces the number of necessary Assisted GPS (AGPS) fixes in the tracking process by utilizing the prior knowledge of an elderly carrier’s location. Such prior knowledge is generated by recogni...[ Read more ]

Recently, societal attention on elderly health care has significantly increased as global aging has accelerated. Much effort has been made to apply Information and Communication Technology (ICT) in the provision of elderly health care services. Elderly tracking, or monitoring the presence information of an elderly person, is one such service. This thesis aims to propose an energy-efficient elderly tracking algorithm. It provides a review of some existing methods on positioning as well as some previous works on people tracking and activity recognition. Our proposed tracking algorithm significantly reduces the number of necessary Assisted GPS (AGPS) fixes in the tracking process by utilizing the prior knowledge of an elderly carrier’s location. Such prior knowledge is generated by recognizing the Personal Common Locations (PCLs) of the elderly carrier based on historic cell ID information. A PCL recognition algorithm and a destination PCL prediction algorithm are included. The proposed tracking algorithm performs energy-conservative tracking without AGPS if prior location knowledge provides sufficient information of the elderly carrier’s current location or movements. Otherwise, it performs AGPS-enabled tracking. We also investigate how to optimize some of the AGPS tracking parameters in order to make our tracking algorithm as energy-efficient as possible. In addition, we describe in detail a tracking system prototype deploying our elderly tracking algorithm. Key components of the system prototype include a cellular phone with High-Speed Downlink Packet Access (HSDPA) and AGPS capability, a centralized tracking server, a tracking database for location information storage, and a web server for data visualization. Tracking experiments using the tracking system prototype are conducted to evaluate our algorithm’s performance with respect to its energy efficiency, its accuracy, and its robustness in different tracking scenarios.