Energy Efficiency in Human Activity Recognition on Smartphones with Adaptive Sampling and Deep Learning Models with Early Exits

Marcel Franco de Oliveira , Alessandro Copetti, Roberto Pacheco , Lauro Alexandre G. Sampaio & Luciano Bertini

Abstract: Human Activity Recognition (HAR) plays a crucial role in applications such as health monitoring, physical training, and therapeutic support. Smartphones equipped with embedded sensors provide a convenient platform for these applications. However, running deep learning models continuously can drain device energy autonomy. Despite recent advances, few studies have explored how executing deep learning-based HAR models across different computational environments affects energy consumption, especially when combined with strategies that leverage application-level contextual information. This work develops and evaluates optimized deep learning models, including variants with early exits, in both local and remote execution scenarios. It also incorporates an adaptive sampling strategy that dynamically adjusts sensor reading frequency according to movement intensity (contextual information). Results show that the proposed approach reduces energy consumption and inference time while preserving model accuracy.

Keywords: Deep learning; early exit; human activity recognition; adaptive sampling; energy efficiency; mobile devices.

DOI code: 10.21528/lnlm-vol24-no1-art1

PDF file: vol24-no1-art1.pdf

BibTex file: vol24-no1-art1.bib