Combining multiple tinyML models for multimodal context-aware stress recognition on constrained microcontrollers

Tools

Woodward, K. ORCID: 0000-0003-3302-1345, Gibbs, M. ORCID: 0009-0009-4754-7242 and Kanjo, E. ORCID: 0000-0002-1720-0661, 2023. Combining multiple tinyML models for multimodal context-aware stress recognition on constrained microcontrollers. Nottingham: Nottingham Trent University.

Preview

Text
1795517_Kanjo.pdf - Published version
Download (1MB) | Preview

Abstract

As stress continues to be a major health concern, there is growing interest in developing effective stress management systems that can detect and mitigate stress in real-world environments. Deep Neural Networks (DNNs) have shown their effectiveness in accurately classifying stress, but most existing solutions rely on the cloud or large obtrusive devices for inference. The emergence of tinyML provides an opportunity to bridge this gap and enable ubiquitous intelligent systems. In this work, we developed a context-aware stress detection approach that uses utilises tinyML to perform continuous inference of physical activity to mitigate motion artifacts when inferring stress from heart rate and electrodermal activity. We explore the challenges and trade-offs in deploying DNNs on resource-constrained microcontrollers for real-world stress recognition. Our proposed context-aware approach can improve the accuracy and privacy of stress detection systems while eliminating the need to store or transmit sensitive health data to the cloud.

Item Type:

Working paper

Creators:

Woodward, K., Gibbs, M. and Kanjo, E.

Publisher:

Nottingham Trent University

Place of Publication:

Nottingham

Date:

26 June 2023

Identifiers: