Abstract:Pattern recognition is the core component of the electronic nose (E-nose). Traditional machine learning algorithms highly rely on the feature data selected manually for model training and testing. A complete experiment must be performed before the data can be further processed. To realize the automatic extraction of response features and simplify the model's training and application process, a multi-task convolutional neural network (MTL-CNN) with a dual-block knowledge-sharing structure is designed to train a model for the E-nose system. This model can simultaneously perform three different classification tasks, for the purposes of target discrimination, concentration prediction, and state judgment. Only a few consecutive seconds of response data are needed to be input into the trained model to obtain various information about the E-nose. With the utilization of an unmanned gas-sensing test system, large-scale measurements of the E-nose can be carried out automatically. A baseline tracking algorithm (BTA) is designed based on the relative changes of short-term data, reducing the impact of long-term shifts. Over thousands of gas response processes and more than 10 million sensing data have participated in the training of the deep learning model. The 5-fold cross-validation method shows that the fully trained model has an outstanding generalization performance. After the baseline is tracked automatically, the accuracy of three tasks towards 12 kinds of volatile organic compounds (VOCs) is about 95% (type recognition: 95.2%, concentration prediction: 92.1%, status judgment: 97.3%) using only 4 seconds of sensing data during the response status of the E-nose. Our work shows the distinct advantages of combining "big data" and "deep learning" in the gas-sensing field and further proves that the application of MTL-CNN can significantly improve the training and application efficiency of the E-nose.

Recent Progress in Smart Electronic Nose Technologies Enabled with Machine Learning Methods

Progress in Bionic Information Processing Techniques for an Electronic Nose Based on Olfactory Models

Recent advances in signal processing algorithms for electronic noses

Pattern Classification Method for Electronic Noses Based on Olfactory Neural Network Using Time Series

Electronic Noses: from Advanced Materials to Sensors Aided with Data Processing

A pattern recognition method for electronic noses based on an olfactory neural network

Electronic Noses: From Gas-Sensitive Components and Practical Applications to Data Processing

Advanced Algorithms for Low Dimensional Metal Oxides-Based Electronic Nose Application: A Review

Electronic Nose for the Detection and Discrimination of Volatile Organic Compounds: Application, Challenges, and Perspectives

Advanced Technology and New Application in Electronic Nose

Recognition Algorithms in E-Nose: A Review

Advances in drift compensation algorithms for electronic nose technology

Target discrimination, concentration prediction, and status judgment of electronic nose system based on large-scale measurement and multi-task deep learning

Smart Electronic Nose Enabled by an All‐Feature Olfactory Algorithm

Application and Progress of Electronic Nose in Environmental Monitoring

A Nonlinear Manifold Embedding Extreme Learning Machine to Improve the Gas Recognition of Electronic Noses

Electronic Nose Feature Extraction Methods: A Review

Multi-task deep learning model for quantitative volatile organic compounds analysis by feature fusion of electronic nose sensing

Sensor array optimization for the electronic nose via different deep learning methods

Efficient Solutions for Discreteness, Drift, and Disturbance (3D) in Electronic Olfaction

Classification techniques of electronic nose: a review