Multi-User Activity Recognition in an Indoor Environment with Transformer Architectures
DOI:
https://doi.org/10.52731/liir.v006.363Keywords:
Human Activity Recognition, HAR, Channel State, CSI, Deep Learning, CNN, TransformerAbstract
This paper proposes a device-free Human Activity Recognition (HAR) system, utilising Wi-Fi Channel State Information (CSI) to maintain the privacy of users in a multi-user environment. To achieve this goal, substantial annotated training data is required, which is often imbalanced with poor generalisability in complex, multi-user environments. To overcome these gaps, a hybrid deep learning approach is proposed that integrates signal pre-processing, targeted data augmentation, and a novel CNN incorporating a Transformer model. Experimental results show that the proposed model outperforms several baselines in single-user and multi-user contexts. Our findings demonstrate that combining real and augmented data significantly improves model generalisation in scenarios with limited labelled data.
References
A. Lotfi, C. Langensiepen, et al., “Smart homes for the elderly dementia sufferers: identification and prediction of abnormal behaviour.” Journal of Ambient Intelligence and Humanized Computing, vol. 3, pp. 205–218 (2012).
H. Chen, et al., “WiFi CSI-Based Activity Recognition Using Attention Mechanisms,” Sensors, vol. 21, no. 2, p. 452, 2021.
A. Naser, A. Lotfi, et. al, ”Privacy-Preserving, Thermal Vision With Human in the Loop Fall Detection Alert System” IEEE Trans. on Human-Machine Systems, Vol. 53, no. 1, pp. 164-175, 2023.
A. Doherty, H. Yuan, et. al, Self-supervised learning of accelerometer data provides new insights for sleep and its association with mortality, 2023.
M. Ezzeldin, S. Ghoneim, et. al, 2024. Multi-modal hybrid hierarchical classification approach with transformers to enhance complex human activity recognition. Signal, Image and Video Processing, 18(12), pp.9375-9385, 2024.
L.Wang, et al., “Privacy Considerations in HAR,” Proc. ACM Interact. Mob.Wearable Ubiquitous Technol., vol. 5, pp. 1–26, 2021.
A.M. Ashleibta, A. Taha, et. al, “5g-enabled contactless multi-user presence and activity detection for independent assisted living,” Sci. Reports, 11(1), p.17590, 2021.
H. Lee, J. Park, and Y. Nam, “Temporal positional encoding for HAR using CSI and transformer architectures,” IEEE Sensors J., vol. 24, no. 3, pp. 4567–4578, 2024.
W. Han and J. Yu, “Algorithm for interference filtering of Wi-Fi gesture recognition,” Int. J. Inf. Commun. Technol., vol. 24, no. 1, pp. 1–20, 2024.
Q. Chen, Y. Lin, and S. Zhao, “Entropy-based features for HAR using CSI,” Sensors, vol. 18, no. 4, pp. 1223–1231, 2018.
