Teach your WiFi-Device: Recognise Gestures and Simultaneous Activities from Time-Domain RF-Features. Sigg, S., Shi, S., & Ji, Y. International Journal on Ambient Computing and Intelligence (IJACI), 2014.
doi  abstract   bibtex   
The authors consider two untackled problems in RF-based activity recognition: the distinction of simultaneously conducted activities of individuals and the recognition of gestures from purely time-domain-based features. Recognition is based on a single antenna system. This is important for the application in end-user devices which are usually single-antenna systems and have seldom access to more sophisticated, e.g. frequency-based features. In case studies with software defined radio nodes utilised in an active, device-free activity recognition DFAR system, the authors observe a good recognition accuracy for the detection of multiple simultaneously conducted activities with two and more receive devices. Four gestures and two baseline situations are distinguished with good accuracy in a second case study.
@article{Sigg_2014_teach,
author={Stephan Sigg and Shuyu Shi and Yusheng Ji},
title={Teach your WiFi-Device: Recognise Gestures and Simultaneous Activities from Time-Domain RF-Features},
journal={International Journal on Ambient Computing and Intelligence (IJACI)},
volume={6},
number={1},
year={2014},
doi={http://dx.doi.org/10.4018/ijaci.2014010102},
abstract={The authors consider two untackled problems in RF-based activity recognition: the distinction of simultaneously conducted activities of individuals and the recognition of gestures from purely time-domain-based features. Recognition is based on a single antenna system. This is important for the application in end-user devices which are usually single-antenna systems and have seldom access to more sophisticated, e.g. frequency-based features. In case studies with software defined radio nodes utilised in an active, device-free activity recognition DFAR system, the authors observe a good recognition accuracy for the detection of multiple simultaneously conducted activities with two and more receive devices. Four gestures and two baseline situations are distinguished with good accuracy in a second case study.},
group = {ambience}}
Downloads: 0