An Approach to Non Invasive Neural Network based Diagnostics of Asthma using Gas Sensors Array

Objectives: This paper presents a neural network based non-invasive diagnostics methodology of asthma using gas sensors array working as artificial olfactory. Methods/Statistical Analysis: A series of invasive clinical trials are recommended by international bodies to justify the diagnosis of asthma. So during these tests the patient has to undergo a lot of physical trauma. To ease the suffering of the patient, in this paper, a non-invasive method for asthma detection has been proposed. Findings: In this paper, an array consisting of five semiconductor gas sensors has been developed. The sensor array along with the data acquisition system has been developed for the non-invasive detection of the asthma. Five data sets of asthma related toxic gas in different concentrations are obtained by a signal acquirement system having tin oxide gas sensor array. Obtained data are put for training and analysis on Artificial Neural Network (ANN). Proposed neural network has been trained using back propagation algorithm. From the results obtained, it can be seen that the developed model ensures the proper accuracy and consistent results. Application/Improvements: Experimental results show good classification of asthma associated exhaled toxic gas with the ambient air using only few samples and also presents the efficiency of Feed Forward Back Propagation Neural Network on the data driven from different gas sensors.