Please use this identifier to cite or link to this item:
|Scopus||Web of Science®||Altmetric|
|Title:||Ambulatory monitoring using passive computational RFID sensors|
|Citation:||IEEE Sensors Journal, 2015; 15(10):5859-5869|
|Asanga Wickramasinghe, and Damith C. Ranasinghe|
|Abstract:||Rapidly emerging batteryless sensors are creating tremendous opportunities for truly wearable sensors for activity recognition. However, data streams from such sensors are characterized by sparsity and noise, which make activity recognition a challenging task. In this paper, we study the feasibility of passive computational RFID sensors for ambulatory monitoring. In particular, we focus on recognizing transfers out of beds or chairs and walking. Ideally, all these activities need to be monitored by movement sensor alarm systems to alert caregivers to provide supervision during the ambulation of older people in hospitals and nursing homes to prevent a fall. Our novel approach to partition continuous sensor data on natural activity boundaries and to identify transfers out of beds or chairs and walking as transitions between sequences of movements overcomes issues posed by the sparsity and the noise. We demonstrate through in-depth experiments the high performance (F-score > 93%) and the responsiveness of our approach.|
|Keywords:||Passive computational RFID sensors; body-worn sensors; activity recognition; ambulatory monitoring; natural activity boundary segmentation|
|Rights:||© 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.|
|Appears in Collections:||Aurora harvest 3|
Computer Science publications
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.