Pattern recognition in integrated sensors. / Reconhecimento de padrões em sensores integrados.

AUTOR(ES)
DATA DE PUBLICAÇÃO

2005

RESUMO

In this work, several methods for pattern recognition and signal processing were studied and applied, using data obtained from different experimental setup of the Electronic Nose. The signals were obtained from the array of conductive sensor into de Nose system, which worked under variable temperature condition. The signal analyses results were used to obtain patterns in order to identify different chemical substances. In addition it was discussed the possibility of the generalization of sensors response, in this sense the response time, sensibility and selectivity of each gas sensors were analyzed. It was discussed using together the signal processing algorithm and pattern recognition process in order to obtain an optimum process of the information extraction and make decision from the data bank. The integrated use of the signal process and pattern recognition promotes the definition and building of the well data banks structures known as DATAWAREHOUSE. These systems will promote the rapid and efficient data manipulation even with high and heterogeneous data banks. It was proposed an information extraction methodology from the noise signal of the gas sensor throughout auto – regressive fitting process together with the Maximum Entropy Method. The Electronic Nose was proposed as consequence of the experimental results, the Nose system proposed contained only one gas sensor. The gas recognition process was made by Fuzzy controller system. This Electronic Nose showed a robust and stable behavior.

ASSUNTO(S)

nariz eletrônico reconhecimento de padrões processamento de sinais signal processing electronic nose pattern recognition

ACESSO AO ARTIGO

Documentos Relacionados