Multi-Frequency Federated Learning for Human Activity Recognition Using Head-Worn Sensors


METADATA ONLY
Loading...

Date

2024

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric
METADATA ONLY

Data

Rights / License

Abstract

Human Activity Recognition (HAR) benefits various application domains, including health and elderly care. Traditional HAR involves constructing pipelines reliant on centralized user data, which can pose privacy concerns as they necessitate the uploading of user data to a centralized server. This work proposes multi-frequency Federated Learning (FL) to enable: (1) privacy-aware ML; (2) joint ML model learning across devices with varying sampling frequency. We focus on head-worn devices (e.g., earbuds and smart glasses), a relatively unexplored domain compared to traditional smartwatch- or smartphone-based HAR. Results have shown improvements on two datasets against frequency-specific approaches, indicating a promising future in the multi-frequency FL-HAR task. The proposed network's implementation is publicly available for further research and development.

Publication status

published

Editor

Book title

2024 International Conference on Intelligent Environments (IE)

Journal / series

Volume

Pages / Article No.

17 - 24

Publisher

IEEE

Event

20th International Conference on Intelligent Environments (IE 2024)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

federated learning; Human activity recognition (HAR); Head-worn sensors; Earables; glasses

Organisational unit

Notes

Funding

Related publications and datasets