Reliability comparison of a fabricated humidity sensor using various artificial intelligence techniques

Abstract

A humidity sensor detects, measures and reports the content of moisture in the air. Using low cost composite materials, a humidity sensor has been fabricated. The characterization has been done using various techniques to prove its surface morphology and working. The fabricated sensor detects relative humidity in the range of 15% to 65%. The life of the sensor has been calculated using different experimental and statistical methods. An expert system has been modeled using different artificial intelligence techniques which predicts failure of the sensor. The Failure prediction of fabricated sensor using Fuzzy Logic, ANN and ANFIS are 81.4%, 97.4% and 98.2% accurate respectively. ANFIS technique proves to be the most accurate technique for prediction of reliability.