Machine learning-based detection of volatile organic compounds in an electronic nose

06/12/2024

Improving the detection and classification of volatile organic compounds (VOCs) requires innovative approaches combining machine learning and advanced electronic nose (e-nose) technologies, such as gas-sensing gels with liquid crystals. The primary objective of this work is to classify various VOCs accurately by implementing Convolutional Neural Networks (CNNs) to capture spatial features and employing Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) for temporal analysis of VOC-gel interactions. The approach employs a dataset of video recordings of VOC interactions with gas-sensing gels under controlled conditions. The dataset, which includes eleven VOCs, was structured into 5 cycles, each exposing the gel to the VOC and recovering to the original morphology. This structure was designed to provide a comprehensive view of the VOC-gel interactions. For the classification task, the first proposed model is a typical CNN architecture, while the second model is based on RNNs with an LSTM layer to introduce the temporal analysis. Results demonstrated the effectiveness of both models, with the first CNN model achieving an accuracy of 94.76% on frame level and model 2 with the LSTM reaching 98.99% accuracy on patch sequences. Cross-validation was also performed to validate the robustness of the models. These results indicate the added value of temporal analysis in VOC detection, significantly improving the model's classification performance. The study concludes that machine learning models can significantly enhance e-nose accuracy and adaptability, especially in applications requiring complex VOC detection.

Autores da comunidade :

• 

  A

  Ana Bárbara Botelho da Silva Correia

  ist177024

Orientadores desta instituição:

• 

  M

  Maria Margarida Campos da Silveira

  ist14026

electrical-engineering-electronic-engineering-information-engineering - Engenharia Eletrotécnica, Eletrónica e Informática

eng - Inglês

Instituto Superior Técnico