Unsupervised deep representation learning to remove motion artifacts in free-mode body sensor networks
Abstract
In body sensor networks, the need to brace sensing devices firmly to the body raises a fundamental barrier to usability. In this paper, we examine the effects of sensing from devices that do not face this mounting limitation. With sensors integrated into common pieces of clothing, we demonstrate that signals in such free-mode body sensor networks are contaminated heavily with motion artifacts leading to mean signal-to-noise ratios (SNRs) as low as -12 dB. Further, we show that motion artifacts at these SNR levels reduce the F1-score of a state-of-the-art algorithm for human-activity recognition by up to 77.1%. In order to mitigate these artifacts, we evaluate the use of statistical (Kalman Filters) and data-driven (Neural Networks) techniques. We show that well-designed methods of representing IMU data with deep neural networks can increase SNRs in free-mode body-sensor networks from -12 dB to +18.2 dB and, as a result, improve the F1-score of recognizing gestures by 14.4% and locomotion activities by 55.3%.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Autoencoders are commonly used to sanitize and denoise raw sensor data, a known problem with wearable signals that impacts the ability to learn patterns in the data.
“Furthermore, autoencoders are commonly used to sanitize and denoise raw sensor data [127,130,148], a known problem with wearable signals that impacts our ability to learn patterns in the data.”
From Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances (2022) · cites this paper as [127]
The DSTSAE outperforms traditional Kalman Filters and improves the SNR from −12 dB to +18.2 dB, with the F1-score of recognizing gestures improved by 14.4% and locomotion activities by 55.3%.
“Experiments show that the DSTSAE outperforms traditional Kalman Filters and improves the SNR from −12 dB to +18.2 dB, with the F1-score of recognizing gestures improved by 14.4% and locomotion activities by 55.3%.”
From Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances (2022) · cites this paper as [127]
Cites
Auto-Encoding Variational Bayes
Kingma, D. P., & Welling, M. (2013) · arXiv · doi:10.48550/arxiv.1312.6114 · Cited by 20,067 · see citing works
Graphical Models, Exponential Families, and Variational Inference
Wainwright, M. J., & Jordan, M. I. (2007) · now publishers, Inc. eBooks · doi:10.1561/9781601981851 · Cited by 3,063 · see citing works
Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition
Ordóñez, F., & Roggen, D. (2016) · Sensors · doi:10.3390/s16010115 · Cited by 2,761 · cited 2× · see citing works
“Among the sensors available, the authors in [9] utilized 113 channels: 3 channels of accelerometer data from 12 locations, 9 channels of IMU data from 5 locations and 16 channels from 2 sensors measuring Euler orientations and angular velocity.”
Graphical Models, Exponential Families, and Variational Inference
Wainwright, M. J., & Jordan, M. I. (2008) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000001 · Cited by 1,842 · see citing works
Pearce, D., & Hirsch, H. (2000) · International Conference on Spoken Language Processing (ICSLP) · doi:10.21437/icslp.2000-743 · Cited by 1,779 · see citing works
Chen, M., González-Valenzuela, S., Vasilakos, A. V., Cao, H., & Leung, V. C. M. (2010) · Mobile Networks and Applications · doi:10.1007/s11036-010-0260-8 · Cited by 1,353 · see citing works
Mathie, M. J., Coster, A. C. F., Lovell, N. H., & Celler, B. G. (2004) · Physiological Measurement · doi:10.1088/0967-3334/25/2/r01 · Cited by 828 · see citing works
The Opportunity challenge: A benchmark database for on-body sensor-based activity recognition
Chavarriaga, R., Sagha, H., Calatroni, A., Digumarti, S. T., Tröster, G., Millán, J. D. R., & Roggen, D. (2013) · Pattern Recognition Letters · doi:10.1016/j.patrec.2012.12.014 · Cited by 742 · cited 1× · see citing works
“DCLSTM outperforms competing HAR algorithms by up to 9%, when applied to the Opportunity dataset [10], [11].”
Enabling technologies for wireless body area networks: A survey and outlook
Cao, H., Leung, V. C. M., Chow, C., & Chan, H. C. B. (2009) · IEEE Communications Magazine · doi:10.1109/mcom.2009.5350373 · Cited by 518 · see citing works
Wireless sensor network based wearable smart shirt for ubiquitous health and activity monitoring
Lee, Y., & Chung, W. (2009) · Sensors and Actuators B Chemical · doi:10.1016/j.snb.2009.04.040 · Cited by 383 · see citing works
Towards the integration of textile sensors in a wireless monitoring suit
Catrysse, M., Puers, R., Hertleer, C., Langenhove, L. V., van Egmond, H., & Matthys, D. (2004) · Sensors and Actuators A Physical · doi:10.1016/j.sna.2003.10.071 · Cited by 325 · see citing works
Long-term monitoring of gait in Parkinson's disease
Moore, S. T., MacDougall, H. G., Graciès, J., Cohen, H. S., & Ondo, W. G. (2007) · Gait & Posture · doi:10.1016/j.gaitpost.2006.09.011 · Cited by 198 · see citing works
Sensor Placement Variations in Wearable Activity Recognition
Kunze, K., & Lukowicz, P. (2014) · IEEE Pervasive Computing · doi:10.1109/mprv.2014.73 · Cited by 111 · cited 1× · see citing works
“One related work is [12], where Kunze and Lukowicz examine the difference between signals acquired from the same sensor when it is placed at different locations on the body (within the context of a cBSN).”
Lapinski, M., Berkson, E. M., Gill, T., Reinold, M., & Paradiso, J. A. (2009) · International Symposium on Wearable Computers · doi:10.1109/iswc.2009.27 · Cited by 79 · see citing works
Analysis of Movement, Orientation and Rotation-Based Sensing for Phone Placement Recognition
İncel, Ö. D. (2015) · Sensors · doi:10.3390/s151025474 · Cited by 69 · see citing works
An Attachable Clothing Sensor System for Measuring Knee Joint Angles
Bergmann, J., Anastasova, S., Spulber, I., Gulati, V., Georgiou, P., & McGregor, A. H. (2013) · IEEE Sensors Journal · doi:10.1109/jsen.2013.2277697 · Cited by 65 · see citing works
Accelerometry-based recognition of the placement sites of a wearable sensor
Mannini, A., Sabatini, A. M., & INTILLE, S. S. (2015) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2015.06.003 · Cited by 50 · see citing works
An overview of body sensor networks in enabling pervasive healthcare and assistive environments
Liolios, C., Doukas, C. N., Fourlas, G. K., et al. (2010) · International Conference on PErvasive Technologies Related to Assistive Environments (PETRA) · doi:10.1145/1839294.1839346 · Cited by 47 · see citing works
Impact of sensor misplacement on estimating metabolic equivalent of task with wearables
Alinia, P., Saeedi, R., Mortazavi, B. J., Rokni, A., & Ghasemzadeh, H. G. (2015) · Wearable and Implantable Body Sensor Networks · doi:10.1109/bsn.2015.7299385 · Cited by 21 · see citing works
Garment Positioning and Drift in Garment-Integrated Wearable Sensing
Gioberto, G., & Dunne, L. E. (2012) · International Symposium on Wearable Computers · doi:10.1109/iswc.2012.35 · Cited by 16 · see citing works
Enokibori, Y., Hayashi, T., & Mase, K. (2015) · doi:10.1145/2800835.2800937 · Cited by 1 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Deep learning for sensor-based activity recognition: A survey
Wang, J., Chen, Y., Hao, S., Peng, X., & Hu, L. (2019) · Pattern Recognition Letters · doi:10.1016/j.patrec.2018.02.010 · Cited by 2,014 · see citing works
Wearable health devices-vital sign monitoring, systems and technologies
Dias, D., & Paulo Silva Cunha, J. (2018) · Sensors · doi:10.3390/s18082414 · Cited by 919 · see citing works
Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances
Zhang, S., Li, Y., Zhang, S., Shahabi, F., Xia, S., Deng, Y., & Alshurafa, N. (2022) · Sensors · doi:10.3390/s22041476 · Cited by 625 · see citing works
“Furthermore, autoencoders are commonly used to sanitize and denoise raw sensor data [98, 190, 280], a known problem with wearable signals that impacts our ability to learn patterns in the data.”
BackgroundMethodValarezo, A., Rivera, L., Park, H., Park, N., & Kim, T.-S. (2020) · Computer Science and Information Systems · doi:10.2298/csis190920005v · Cited by 12 · see citing works
- Authors
- Mohammed, S., & Tashev, I.
- Affiliations
- Microsoft (United States) (US)
- ORCID
- Tashev, Ivan
- Venue
- 2017 IEEE 14th International Conference on Wearable and Implantable Body Sensor Networks (BSN)
- Publisher
- IEEE
- Published
- 2017-05
- Pages
- 183-188
- ISSN
- 2376-8886 (ISSN-L)
- OpenAlex
- W2605049121
- Cited by
- 29
- References
- 21
- Citations by year
- 2026: 1 · 2025: 1 · 2024: 2 · 2023: 2 · 2022: 3 · 2021: 5 · 2020: 4 · 2019: 2 · 2018: 9
- FWCI
- 1.20
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Wearable and Implantable Body Sensor Networks (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Physical Sciences › Engineering › Biomedical Engineering › Wireless Body Area Networks
- Physical Sciences › Engineering › Biomedical Engineering › Non-Invasive Vital Sign Monitoring
- Keywords
- motion artifact removal, wearable sensors, IMU data, human activity recognition, gesture recognition, signal-to-noise ratio, Kalman filter, artificial neural networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
