A Theory for Multiresolution Signal Decomposition: The Wavelet Representation
IEEE Transactions on Pattern Analysis and Machine Intelligence
| Hi-index | 0.00 |
Calibrated identification of volatile organics by electronic sensors needs development of data collection and data processing methods that can efficiently generate vapor identity features and some quantitative measure of its concentration simultaneously. In this paper, we present a simulation study on this based on surface acoustic wave (SAW) chemical sensors functionalized by polymer coating. The analysis utilizes transient responses of SAW sensors exposed to seven volatile organic compounds at various concentrations. The feature extraction is done by discrete wavelet decomposition using Daubechies-2 basis. A fuzzy c-means clustering method based Sugeno-type fuzzy inference system was then roped in for simultaneous identification and concentration estimation. The performance of the method has been analyzed for various conditions of polymer film thickness. It is concluded that there exists an optimum region for film thickness over which the present method yields nearly 100% correct classification with less than 1% concentration error.