Summary
Asia-Pacific Network Operations and Management Symposium
2016
Session Number:P1
Session:
Number:P1-30
RTagCare: Deep Human Activity Recognition Powered by Passive Computational RFID Sensors
Guibing Hu, Xuesong Qiu, Meng Luoming,
pp.-
Publication Date:2016/10/5
Online ISSN:2188-5079
DOI:10.34385/proc.25.P1-30
PDF download (670.2KB)
Summary:
Activity recognition is a hot topic of research that is widely adopted by many applications such as fall detection of elderly people. Emerging passive RFID (radio-frequency identification) is creating huge opportunity for wearable devices to achieve activity recognition. However, performance of activity recognition is constrained by RFID localization accuracy and low quality of data streams characterized by sparsity and noise. In this paper, we present a novel activity recognition system, called RTagCare, which is a low-cost, unobtrusive and lightweight RFID based system. The RTagCare system leverage RFID localization technology, 3D-accelerometer base human activity identification and data mining algorithm to overcome traditional activity recognition system issues. RTagCare has been implemented and deployed in a test environment. As a result, RTagCare generally performs well to recognize human activity with high performance (F-score >94%).