Electronic Nose Feature Extraction Methods: A Review

Abstract

Many research groups in academia and industry are focusing on the performance improvement of electronic nose (E-nose) systems mainly involving three optimizations, which are sensitive material selection and sensor array optimization, enhanced feature extraction methods and pattern recognition method selection. For a specific application, the feature extraction method is a basic part of these three optimizations and a key point in E-nose system performance improvement. The aim of a feature extraction method is to extract robust information from the sensor response with less redundancy to ensure the effectiveness of the subsequent pattern recognition algorithm. Many kinds of feature extraction methods have been used in E-nose applications, such as extraction from the original response curves, curve fitting parameters, transform domains, phase space (PS) and dynamic moments (DM), parallel factor analysis (PARAFAC), energy vector (EV), power density spectrum (PSD), window time slicing (WTS) and moving window time slicing (MWTS), moving window function capture (MWFC), etc. The object of this review is to provide a summary of the various feature extraction methods used in E-noses in recent years, as well as to give some suggestions and new inspiration to propose more effective feature extraction methods for the development of E-nose technology.

Ripening season, ethylene production and respiration rate are related to fruit non-destructively-analyzed volatiles measured by an electronic nose in 57 peach (Prunus persica L.) samples

“…E-noses have non-linearity characteristics of sensor response to the odors called "fingerprint" (Yan et al, 2015). In our study, intensity values of T30/1, T70/2, PA/2, P30/1, P40/2, P30/2 and T40/2 sensors were in accord with all the three classifications of harvest season (ethylene and respiration, too), respiration rate and ethylene production ( Fig.…”

Highly Sensitive Sensors Based on Metal-Oxide Nanocolumns for Fire Detection

“…When highly sensitive sensors respond to target gases, their resistance drops to 1/10,000 from 1/1000 compared to their resistance in air ambient. Thus, before decision making process, we convert the dynamic ranges of the sensors comparable by unity-based normalized data of each sensor to the interval [0,1] using the following linear transformation [34,35]: $s^{\prime }\left[i\right]=\frac{s\left[i\right]-S_{min}}{S_{max}-S_{min}},0\le i\le n,$ where $s^{\prime }\left[i\right]$ is the normalized sensor response and $S_{min}$, $S_{max}$ are minimum, maximum value of filtered data among $s\left[0\right]$, …, $s\left[n\right]$, respectively. In most cases, $S_{max}$ indicates baseline sensor resistance where sensor is in air ambient and $S_{min}$ indicates sensor resistance exposed to maximum concentration of toxic gases.…”

Flexible Multifunctional Sensors for Wearable and Robotic Applications

“…It has a temperature sensitivity of 4.05 kH °C −1 with a sensing range from 25 to 95 °C and pressure sensitivity of 8.19 kHz bar −1 for 0-5 bar. The novel electronic-nose (e-nose) system [107] is specifically used to sense molecules in analogy to the human nose. Based on longitudinal and shear mode resonance, a dual-mode liquid sensor was developed with c-axis tilted piezoelectric aluminum nitride (AlN) film.…”

“…The mass sensitivity of this sensor achieved a value up to 1.72 × 10 3 Hz cm 2 ng −1 . The novel electronic-nose (e-nose) system [107] is specifically used to sense molecules in analogy to the human nose. It can discriminate various volatile organic compounds (VOC) and quantitate compound concentrations and could be developed with a multimode FBAR technique.…”