Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
2 files | 1.41 MB
Add to list
  1. Search results
  2. Quantifying uncertainty in real time ...

Quantifying uncertainty in real time with split BiRNN for radar human activity recognition

Lorin Werthen-Brabants (UGent) , Geethika Bhavanasi, Ivo Couckuyt (UGent) , Tom Dhaene (UGent) and Dirk Deschrijver (UGent)
(2022) 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD). In European Radar Conference EuRAD p.173-176
Author
Organization
Project
Abstract
Radar systems can be used to perform human activity recognition in a privacy preserving manner. Deep Neural Networks are able to effectively process the complex radar data and make predictions. Often these networks are large and do not scale well when processing a large amount of radar streams at once, for example when monitoring multiple rooms in a hospital. This work proposes Bayesian Split Bidirectional Recurrent Neural Network for Human Activity Recognition. Using this technique the processing of data is split in two parts, one part on-premise (low-power, low-cost device), and the other off-premise (high power device). The proposed approach leverages the power of the off-premise device to quantify its uncertainty, and to gain more information on its epistemic and its aleatoric parts. Results indicate the proposed approach is able to correctly identify parts of a prediction that either need more training data for better predictions (epistemic uncertainty), or are inherently hard to classify by the model (aleatoric uncertainty).
Keywords
bayesian neural networks, uncertainty quantification, edge-cloud interaction, bidirectional rnn, human activity recognition, radar

Downloads

  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 705.04 KB
  • Download
    eurad split birnn-1.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 705.40 KB

Citation

Please use this url to cite or link to this publication:

MLA
Werthen-Brabants, Lorin, et al. “Quantifying Uncertainty in Real Time with Split BiRNN for Radar Human Activity Recognition.” 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), IEEE, 2022, pp. 173–76, doi:10.23919/eurad54643.2022.9924932.
APA
Werthen-Brabants, L., Bhavanasi, G., Couckuyt, I., Dhaene, T., & Deschrijver, D. (2022). Quantifying uncertainty in real time with split BiRNN for radar human activity recognition. 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), 173–176. https://doi.org/10.23919/eurad54643.2022.9924932
Chicago author-date
Werthen-Brabants, Lorin, Geethika Bhavanasi, Ivo Couckuyt, Tom Dhaene, and Dirk Deschrijver. 2022. “Quantifying Uncertainty in Real Time with Split BiRNN for Radar Human Activity Recognition.” In 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), 173–76. IEEE. https://doi.org/10.23919/eurad54643.2022.9924932.
Chicago author-date (all authors)
Werthen-Brabants, Lorin, Geethika Bhavanasi, Ivo Couckuyt, Tom Dhaene, and Dirk Deschrijver. 2022. “Quantifying Uncertainty in Real Time with Split BiRNN for Radar Human Activity Recognition.” In 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), 173–176. IEEE. doi:10.23919/eurad54643.2022.9924932.
Vancouver
1.
Werthen-Brabants L, Bhavanasi G, Couckuyt I, Dhaene T, Deschrijver D. Quantifying uncertainty in real time with split BiRNN for radar human activity recognition. In: 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD). IEEE; 2022. p. 173–6.
IEEE
[1]
L. Werthen-Brabants, G. Bhavanasi, I. Couckuyt, T. Dhaene, and D. Deschrijver, “Quantifying uncertainty in real time with split BiRNN for radar human activity recognition,” in 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), Milan, Italy, 2022, pp. 173–176.
@inproceedings{8773201,
  abstract     = {{Radar systems can be used to perform human activity recognition in a privacy preserving manner. Deep Neural Networks are able to effectively process the complex radar data and make predictions. Often these networks are large and do not scale well when processing a large amount of radar streams at once, for example when monitoring multiple rooms in a hospital. This work proposes Bayesian Split Bidirectional Recurrent Neural Network for Human Activity Recognition. Using this technique the processing of data is split in two parts, one part on-premise (low-power, low-cost device), and the other off-premise (high power device). The proposed approach leverages the power of the off-premise device to quantify its uncertainty, and to gain more information on its epistemic and its aleatoric parts. Results indicate the proposed approach is able to correctly identify parts of a prediction that either need more training data for better predictions (epistemic uncertainty), or are inherently hard to classify by the model (aleatoric uncertainty).}},
  author       = {{Werthen-Brabants, Lorin and Bhavanasi, Geethika and Couckuyt, Ivo and Dhaene, Tom and Deschrijver, Dirk}},
  booktitle    = {{2022 19TH EUROPEAN RADAR CONFERENCE (EURAD)}},
  isbn         = {{9782874870712}},
  keywords     = {{bayesian neural networks,uncertainty quantification,edge-cloud interaction,bidirectional rnn,human activity recognition,radar}},
  language     = {{eng}},
  location     = {{Milan, Italy}},
  pages        = {{173--176}},
  publisher    = {{IEEE}},
  title        = {{Quantifying uncertainty in real time with split BiRNN for radar human activity recognition}},
  url          = {{http://doi.org/10.23919/eurad54643.2022.9924932}},
  year         = {{2022}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: