Font Size: a A A

Research On Smartongue Based On Sinusoidal Envelope Signal Technique And Application

Posted on:2013-02-12Degree:MasterType:Thesis
Country:ChinaCandidate:X XiaoFull Text:PDF
GTID:2211330371468141Subject:Food Science and Engineering
Abstract/Summary:PDF Full Text Request
Smartongue system is a kind of voltammetric electronic tongue self-developed in our laboratory, which composed of sensor array, signal excitation and acquisition system and intelligent algorithm. A novel stimulating signal named sinusoidal envelope signal was designed in this research based on the experimental platform for the Smartongue hardware. Meanwhile, the sinusoidal envelope signal was verified in capability and stability detections comparing to multi-frequency large amplitude pulse voltage, after response data process and analysis. And it has successful application to Chinese white spirits separation, Chinese rice wines from different areas detection, coconut milk character discrimination and pasteurized milk shelf-life identification. The main research works and results are as follows:(1) This research has created sinusoidal envelope detecting signal consisted1Hz,10Hz and100Hz frequency segments detected envelope signal, of which waveform was optimized related to the effects of four taste substances separation. Based on the construction of Smartongue experimental platform, the sinusoidal envelope signal detection techniques has been established combining with an array of non-modified metal electrode sensors and multivariate statistical analysis.(2) Considering to the characteristics of sinusoidal envelope response signals, we chose the principal component analysis and Fourier transform as data processing methods, which were accomplished by MATLAB software. Comparative study of the real-time and off-line Fourier transform processing had been carried out to select off-line Fourier transform as data processing methods. Based on analysis of frequency-amplitude spectra made from sinusoidal envelope response data after fast Fourier transformation, we established an effective analysis method, named Fourier frequency-amplitude fingerprint. The analysis results pointed out that recognition ability on different substances of the Smartongue system was sinusoidal envelope signal detection based on the wisdom of the tongue system is reflected from the differences of amplitude values on frequency-amplitude spectra.(3) This research was willing to explore signal envelope signal technology's detecting capacity and stability using principal component score plots and DI value as evaluation indexes. Compared with multi-frequency large amplitude pulse voltage signal, the results showed that the sinusoidal envelope signal had a better detecting ability and stability of better detection without doubt.(4) Sinusoidal envelope signal had a good application in distinction of Chinese white spirits with different flavor and could classify the specific sensory properties. It could also identified Hua-Diao and Jia-Fan rice wines with different areas and brands in order to play a positive role in origin protection. It had been well applied in coconut milk and distinguished sensory attributes differences and had notable effects on commodities'brand protection and fakes recognition. It could also detect pasteurized milk shelf-life and make judgments about the abnormal milk to warn the tasks. We concluded that sinusoidal envelope signal have a better detection capability than multi-frequency large amplitude pulse voltage signal considering to comparative research on various applications.
Keywords/Search Tags:Smartongue, Sinusoidal envelope signal, Fourier transform, Detection capability, Stability
PDF Full Text Request
Related items