Electronic nose is mainly composed of gas sensor array, signal preprocessing and pattern recognition. An odor appears in front of an active material sensor, and the sensor converts the chemical input into an electrical signal, and the response spectrum of the sensor array to the odor is formed by the responses of multiple sensors to the odor. Obviously, all kinds of chemical components in the smell will interact with sensitive materials, so this response spectrum is a broad-spectrum response spectrum of the smell. In order to realize the qualitative or quantitative analysis of odor, the sensor signal must be properly preprocessed (noise elimination, feature extraction, signal amplification, etc.) and then processed by appropriate pattern recognition analysis methods. Theoretically, every smell has its characteristic response spectrum, and different smells can be distinguished according to its characteristic response spectrum. At the same time, the cross sensitivity of various gases can be measured by using an array of gas sensors, and the analysis of mixed gases can be realized by appropriate analysis methods.
electronic nose uses the characteristic that each gas sensor has different responses to complex gases, and uses data processing methods to identify various odors, so as to analyze and evaluate the odor quality.
The main mechanism of electronic nose recognition is that each sensor in the array has different sensitivity to the gas to be measured. For example, gas No.1 can produce high response on one sensor and low response on other sensors; Similarly, the sensor with high response of No.2 gas is not sensitive to No.1 gas. In the final analysis, the response pattern of the whole sensor array to different gases is different. It is this difference that enables the system to identify gases according to the response pattern of the sensor.
The work of electronic nose can be simply summarized as: sensor array-signal preprocessing-neural network and various algorithms-computer identification (gas qualitative and quantitative analysis). Functionally speaking, gas sensor array is equivalent to a large number of olfactory receptor cells in biological olfactory system, neural network and computer recognition are equivalent to biological brain, and the rest are equivalent to olfactory nerve signal transmission system.