Federated learning with multi-layer perceptron for sensor drift compensation in direct-ethanol fuel cells

Thumbnail Image

Files

master_Pavlyuk_Alexander_2025.pdf (13.96 MB)

URL

Journal Title

Journal ISSN

Volume Title

School of Science | Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.

Authors

Date

2025-07-02

Department

Major/Subject

Machine Learning, Data Science and Artificial Intelligence

Mcode

Degree programme

Master's Programme in Computer, Communication and Information Sciences

Language

en

Pages

67

Series

Abstract

Accurate measurement of breath alcohol concentration (BrAC) is essential for safety-critical applications such as traffic law enforcement. However, modern breathalyzers lack self-diagnosis functionality and their measurement accuracy cannot be assessed on site. Furthermore, the calibration interval is highly dependent on the operating conditions. As a result, a breathalyzer may become inaccurate long before its scheduled calibration. This project proposes a novel method to estimate the accuracy of the measurement of a breathalyzer based on its voltage decay curve obtained under operating conditions. A federated learning (FL) approach is used to address the heterogeneity and privacy of datasets collected from individual breathalyzers. The results demonstrate significant potential for accurately estimating the accuracy of the measurement of breathalyzers.

Description

Supervisor

Jung, Alex

Keywords

federated learning, machine learning, multi-layer perceptron (MLP), sensor drift compensation, fuel cells, direct-ethanol fuel cell (DEFC)

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202508196560

Collections

[dipl] Perustieteiden korkeakoulu / SCI