Arquitetura de Aprendizado Federado para Wearables: Detecção de Risco com FC e HRV na Indústria
Resumo
Monitoring physiological stress in industrial and mining workers is critical for occupational safety. Wearable devices enable real-time measurement of heart rate (HR) and heart rate variability (HRV), but data privacy regulations impose severe restrictions. This work proposes a federated learning (FL) architecture where wearable devices train local models using HR and HRV, adjusted by the worker’s age. Only model parameters are shared with a central server, preserving privacy. Alerts are issued locally when HR exceeds 85% of the age-adjusted maximum heart rate (FCmax = 208 − 0.7 × age) or HRV drops below 20 ms. A proof-of-concept implementation demonstrates the feasibility of decentralized health risk detection using real medical thresholds, without exposing raw physiological data.
Palavras-chave:
Aprendizado Federado, Wearables, Indústria 4.0, Privacidade, Hiperautomação, aprendizado de máquina, automação adaptativa
Referências
Norma Regulamentadora NR1 (Portaria 1.419/24), Ministério do Trabalho e Emprego, Brasil, 2024.
Norma Regulamentadora NR7, Programa de Controle Médico de Saúde Ocupacional (PCMSO), Ministério do Trabalho e Emprego, Brasil.
Norma Regulamentadora NR9, Avaliação e Controle das Exposições Ocupacionais a Agentes Físicos, Químicos e Biológicos, Ministério do Trabalho e Emprego, Brasil.
Norma Regulamentadora NR15, Atividades e Operações Insalubres, Ministério do Trabalho e Emprego, Brasil.
P. P. Janani et al., ”Wearable sensors for physiological monitoring of industrial workers: a review,”Sensors (Basel), vol. 23, no. 18, p. 7709, 2023.
H. Tanaka et al., ”Age-predicted maximal heart rate revisited,”J. Am. Coll. Cardiol., vol. 60, no. 12, pp. 1109–1110, 2012.
S. Laborde et al., ”Heart rate variability and cardiac vagal tone in psychophysiological research: recommendations for experiment planning, data analysis, and reporting,”Front. Psychol., vol. 8, p. 213, 2017.
S. Das et al., ”Federated learning for personalized healthcare: A privacy-preserving and efficient approach,”IEEE J. Biomed. Health Inform., 2024 (in press).
H. Aouedi et al., ”Federated Learning: A Comprehensive Survey on Communication Efficiency, Attacks, and Applications in Healthcare and Industry 4.0,”J. Ambient Intell. Humaniz. Comput., 2024.
B. McMahan et al., ”Communication-Efficient Learning of Deep Networks from Decentralized Data,”Proc. AISTATS, 2017.
P. Marada et al., ”FedAvg and Beyond: A Review of Aggregation Algorithms in Federated Learning,”Proc. IEEE Conf. AI, 2025 (in press).
S. Wang et al., ”Federated Learning for Non-IID Data: A Comprehensive Survey,”ACM Comput. Surv., vol. 55, no. 7, pp. 1–36, 2023.
Apple Inc., ”Apple Watch Series 8: Health and Safety Features,”2023.
Google LLC, ”Fitbit Sense 2: Advanced Health Monitoring,”2023.
F. Shaffer and J. P. Ginsberg, ”An overview of heart rate variability metrics and norms,”Front. Public Health, vol. 5, p. 258, 2017.
W. L. Kenney et al., ”American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults,”Med. Sci. Sports Exerc., vol. 43, no. 3, pp. 560–562, 2011.
C. Montini et al., ”Wearable devices for industrial applications: a selection methodology based on machine learning requirements,”Sensors (Basel), vol. 22, no. 5, p. 1957, 2022.
V. Warrier et al., ”Privacy-Preserving Stress Detection using Federated Learning on Wearable Devices,”Proc. ACM CHASE, 2024.
M. F. Jorge et al., ”Comparing Centralized and Federated Learning for Freezing of Gait Detection from Wearable Sensor Data,”IEEE J. Biomed. Health Inform., 2024 (in press).
Norma Regulamentadora NR7, Programa de Controle Médico de Saúde Ocupacional (PCMSO), Ministério do Trabalho e Emprego, Brasil.
Norma Regulamentadora NR9, Avaliação e Controle das Exposições Ocupacionais a Agentes Físicos, Químicos e Biológicos, Ministério do Trabalho e Emprego, Brasil.
Norma Regulamentadora NR15, Atividades e Operações Insalubres, Ministério do Trabalho e Emprego, Brasil.
P. P. Janani et al., ”Wearable sensors for physiological monitoring of industrial workers: a review,”Sensors (Basel), vol. 23, no. 18, p. 7709, 2023.
H. Tanaka et al., ”Age-predicted maximal heart rate revisited,”J. Am. Coll. Cardiol., vol. 60, no. 12, pp. 1109–1110, 2012.
S. Laborde et al., ”Heart rate variability and cardiac vagal tone in psychophysiological research: recommendations for experiment planning, data analysis, and reporting,”Front. Psychol., vol. 8, p. 213, 2017.
S. Das et al., ”Federated learning for personalized healthcare: A privacy-preserving and efficient approach,”IEEE J. Biomed. Health Inform., 2024 (in press).
H. Aouedi et al., ”Federated Learning: A Comprehensive Survey on Communication Efficiency, Attacks, and Applications in Healthcare and Industry 4.0,”J. Ambient Intell. Humaniz. Comput., 2024.
B. McMahan et al., ”Communication-Efficient Learning of Deep Networks from Decentralized Data,”Proc. AISTATS, 2017.
P. Marada et al., ”FedAvg and Beyond: A Review of Aggregation Algorithms in Federated Learning,”Proc. IEEE Conf. AI, 2025 (in press).
S. Wang et al., ”Federated Learning for Non-IID Data: A Comprehensive Survey,”ACM Comput. Surv., vol. 55, no. 7, pp. 1–36, 2023.
Apple Inc., ”Apple Watch Series 8: Health and Safety Features,”2023.
Google LLC, ”Fitbit Sense 2: Advanced Health Monitoring,”2023.
F. Shaffer and J. P. Ginsberg, ”An overview of heart rate variability metrics and norms,”Front. Public Health, vol. 5, p. 258, 2017.
W. L. Kenney et al., ”American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults,”Med. Sci. Sports Exerc., vol. 43, no. 3, pp. 560–562, 2011.
C. Montini et al., ”Wearable devices for industrial applications: a selection methodology based on machine learning requirements,”Sensors (Basel), vol. 22, no. 5, p. 1957, 2022.
V. Warrier et al., ”Privacy-Preserving Stress Detection using Federated Learning on Wearable Devices,”Proc. ACM CHASE, 2024.
M. F. Jorge et al., ”Comparing Centralized and Federated Learning for Freezing of Gait Detection from Wearable Sensor Data,”IEEE J. Biomed. Health Inform., 2024 (in press).
Publicado
24/11/2025
Como Citar
CAMPOS, Bruno; OLIVEIRA, Ricardo Augusto Rabelo.
Arquitetura de Aprendizado Federado para Wearables: Detecção de Risco com FC e HRV na Indústria. In: ARTIGOS COMPLETOS - SIMPÓSIO BRASILEIRO DE ENGENHARIA DE SISTEMAS COMPUTACIONAIS (SBESC), 15. , 2025, Campinas/SP.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 19-24.
ISSN 2763-9002.
DOI: https://doi.org/10.5753/sbesc_estendido.2025.15331.
