Enhancement of the neural network modeling accuracy using a submodeling decomposition-based technique, application in gas sensor

This paper presents an algorithm based on the use of artificial neural networks (ANNs) in order to reduce the processing time and to improve the accuracy in ANN modeling, which can be accomplished with a division of the model to submodels by input subintervals. We apply this method with a gas sensor aiming to accurately control the small gas leaks, thus decreasing the risk of false alarms and missed detections. The sensor model accurately, especially in small concentrations, expresses the nonlinear character of the response and the dependence on temperature and relative humidity in addition to the gas nature dependency. The corrector linearizes and compensates the sensor’s responses. The results obtained show the effectiveness of the proposed technique.