A Transformer Model-Based Methodology for Person-Independent Human Activity Recognition Using Wi-Fi CSI Data

• Allan Costa Nascimento dos Santos UFF / Brunel University London http://orcid.org/0000-0002-9400-5990
• Iandra Galdino UFF https://orcid.org/0000-0002-3564-2092
• Julio C. H. Soto UFF https://orcid.org/0000-0003-1439-762X
• Taiane C. Ramos UFF https://orcid.org/0000-0002-0351-1450
• Celio V. N. de Albuquerque UFF https://orcid.org/0000-0002-7959-6569
• Raphael Guerra UFF https://orcid.org/0000-0002-6371-0662
• Cledson de Sousa UFF https://orcid.org/0000-0003-3892-1799
• Natalia C. Fernandes UFF https://orcid.org/0000-0001-9481-6374
• Débora Muchaluat-Saade UFF https://orcid.org/0000-0002-1233-9736
• Gheorghita Ghinea Brunel University London https://orcid.org/0000-0003-2578-5580

DOI: https://doi.org/10.5753/sbrc.2025.6443

Abstract

By capturing and interpreting Wi-Fi signals in indoor environments, CSI can be used to detect physical activity, falls, or daily movements of a patient, allowing caregivers and healthcare professionals to monitor patients without the need for wearable sensors or invasive cameras. Therefore, this paper proposes a methodology called MPA-CSI to identify the activity of a person in a room through the analysis of CSI data and a dataset used for its evaluation. MPA-CSI uses Transformer models developed to process time series data featuring a structure that allows capturing temporal dependencies. MPA-CSI is capable of identifying activities of people who did not participate in the training phase. The movement identification accuracy is 96.67% using a dataset with CSI data from 59 volunteers.

Keywords: Transformer model, Channel state information, Wi-Fi, Wireless sensing system, Healthcare, Human activity recognition

References

Caballero, E., Galdino, I., Soto, J. C., Ramos, T. C., Guerra, R., Muchaluat-Saade, D., & Albuquerque, C. (2023). Human activity recognition using Wi-Fi CSI. In International Conference on Pervasive Computing Technologies for Healthcare (pp. 309–321). Springer.

Chen, Z., Zhang, L., Jiang, C., Cao, Z., & Cui, W. (2019). Wi-Fi CSI based passive human activity recognition using attention-based BLSTM. IEEE Transactions on Mobile Computing, 18(11), 2714–2724.

de Sousa, C., Fernandes, V., Coimbra, E. A., & Huguenin, L. (2024). Subcarrier selection for HAR using CSI and CNN: Reducing complexity and enhancing accuracy. In Proceedings of the IEEE Virtual Conference on Communications. Accepted for publication.

dos Santos, A. C. N., de Paula, K., Vidal, M. T. L., da Silva, J. M. M., de Sousa, C., Fernandes, L. A. F., de Castro, T. B., Bedo, M., Kohwalter, T. C., Bastos, C. A. M., Seixas, F. L., Fernandes, N. C., Muchaluat-Saade, D. C., & Ghinea, G. (2024). A computer vision model to support individuals with disabilities within university campuses. In 2024 IEEE International Conference on E-health Networking, Application & Services (HealthCom) (pp. 1–7).

dos Santos, A. C. N., Seixas, F. L., & Fernandes, N. C. (2022). Provendo um modelo automático de detecção de quedas baseado em rede adversária generativa para assistência de idosos. In Anais do XXII Simpósio Brasileiro de Computação Aplicada à Saúde (pp. 120–131). SBC.

Galdino, I., Soto, J. C. H., Caballero, E., Ferreira, V., Ramos, T. C., Albuquerque, C., & Muchaluat-Saade, D. C. (2023). eHealth CSI: A Wi-Fi CSI dataset of human activities. IEEE Access, 11, 71003–71012.

Gouveia, B. G., Galdino, I., Caballero, E., Soto, J. C. H., Ramos, T. C., Guerra, R., Muchaluat-Saade, D., & Albuquerque, C. V. N. (2024). Parameter tuning for accurate heart rate measurement using Wi-Fi signals. In 2024 International Conference on Computing, Networking and Communications (ICNC) (pp. 407–411).

Lee, S., Park, Y. D., Suh, Y. J., & Jeon, S. (2018). Design and implementation of monitoring system for breathing and heart rate pattern using Wi-Fi signals. In IEEE Annual Consumer Communications and Networking Conference (pp. 1–7).

Li, S., Ge, Y., Shentu, M., Zhu, S., Imran, M., Abbasi, Q., & Cooper, J. (2021). Human activity recognition based on collaboration of vision and Wi-Fi signals. In 2021 International Conference on UK-China Emerging Technologies (UCET) (pp. 204–208).

Rothman, D. (2021). Transformers for natural language processing: Build innovative deep neural network architectures for NLP with Python, PyTorch, TensorFlow, BERT, RoBERTa, and more. Packt Publishing Ltd.

Santos, A. C., Firmino, R. M., Soto, J. C., Medeiros, D. S., Mattos, D. M., Albuquerque, C. V., Seixas, F., Muchaluat-Saade, D. C., & Fernandes, N. C. (2020). Aplicações em redes de sensores na área da saúde e gerenciamento de dados médicos: Tecnologias em ascensão. Sociedade Brasileira de Computação.

Soto, J. C., Galdino, I., Caballero, E., Ferreira, V., Muchaluat-Saade, D., & Albuquerque, C. (2022a). A survey on vital signs monitoring based on Wi-Fi CSI data. Computer Communications, 195, 99–110.

Soto, J. C. H., Galdino, I., Gouveia, B. G., Caballero, E., Ferreira, V., Muchaluat-Saade, D., & Albuquerque, C. (2022b). Wi-Fi CSI-based human presence detection using DTW features and machine learning. In 2022 IEEE Latin-American Conference on Communications (LATINCOM) (pp. 1–6).

Wang, D., Yang, J., Cui, W., Xie, L., & Sun, S. (2021). Multimodal CSI-based human activity recognition using GANs. IEEE Internet of Things Journal, 8(24), 17345–17355.

Wang, Y., Wu, K., & Ni, L. M. (2017). WiFall: Device-free fall detection by wireless networks. IEEE Transactions on Mobile Computing, 16(2), 581–594.

Xiao, C., Han, D., Ma, Y., & Qin, Z. (2019). CSI-GAN: Robust channel state information-based activity recognition with GANs. IEEE Internet of Things Journal, 6(6), 10191–10204.