Learning-Based Odor Anomaly Detection Using Bosch BME688 in Indoor Environments

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26220%2F26%3A0201198" target="_blank" >RIV/00216305:26220/26:0201198 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1109/ICUMT67815.2025.11268723" target="_blank" >http://dx.doi.org/10.1109/ICUMT67815.2025.11268723</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICUMT67815.2025.11268723" target="_blank" >10.1109/ICUMT67815.2025.11268723</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Learning-Based Odor Anomaly Detection Using Bosch BME688 in Indoor Environments

Popis výsledku v původním jazyce

This paper presents a complete pipeline for odorbased anomaly detection using the Bosch BME688 metal-oxide gas sensor in indoor environments. Measurements were conducted across several classes of everyday and chemically aggressive odors, including Normal Air, Coffee, Vinegar, Acetone, Bleach, Biowaste, and Paint Thinner. The acquired resistance data were preprocessed, normalized, and analyzed using dimensionality reduction techniques (UMAP, PCA) to explore data structure and class separability.Three anomaly detection models were implemented and compared: a Long Short-Term Memory (LSTM) autoencoder, a dense (fully connected) autoencoder, and the Isolation Forest algorithm. Results show that while the dense autoencoder achieved the highest classification accuracy, particularly on borderline classes, the LSTM autoencoder provided better anomaly separation for well-defined anomaly classes. The findings demonstrate that deep learning models based on resistance signals from the BME688 can effectively detect odor anomalies.

Název v anglickém jazyce

Learning-Based Odor Anomaly Detection Using Bosch BME688 in Indoor Environments

Popis výsledku anglicky

This paper presents a complete pipeline for odorbased anomaly detection using the Bosch BME688 metal-oxide gas sensor in indoor environments. Measurements were conducted across several classes of everyday and chemically aggressive odors, including Normal Air, Coffee, Vinegar, Acetone, Bleach, Biowaste, and Paint Thinner. The acquired resistance data were preprocessed, normalized, and analyzed using dimensionality reduction techniques (UMAP, PCA) to explore data structure and class separability.Three anomaly detection models were implemented and compared: a Long Short-Term Memory (LSTM) autoencoder, a dense (fully connected) autoencoder, and the Isolation Forest algorithm. Results show that while the dense autoencoder achieved the highest classification accuracy, particularly on borderline classes, the LSTM autoencoder provided better anomaly separation for well-defined anomaly classes. The findings demonstrate that deep learning models based on resistance signals from the BME688 can effectively detect odor anomalies.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20206 - Computer hardware and architecture

Projekt

<a href="/cs/project/FW07010015" target="_blank" >FW07010015: Monitorování kvality vnitřního prostředí budov pomocí senzorů pachu a umělé inteligence </a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

2025 17th International Congress on Ultra Modern Telecommunications and Control Systems and Workshops (ICUMT)

ISBN

979-8-3315-7675-2

ISSN

—

e-ISSN

—

Počet stran výsledku

6

Strana od-do

176-181

Název nakladatele

—

Místo vydání

—

Místo konání akce

Florencie, Itálie

Datum konání akce

3. 11. 2025

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—