Neuro-Fuzzy for Sensor Fault Detection and Isolation

Abstract

This paper presents the sensor fault configuration through Neuro-Fuzzy. As we know that sensor faults have been observed in may domain. Various sensors faults are present such as bias, scaling, drift so to remove this kind of fault which is present we make the sensor to reconfigure to normal condition and this reconfiguration is done through Neuro-Fuzzy which uses the expert knowledge stored in them while training. This technique is implemented through ANFIS tool. Sugeno-Type fuzzy inference system is used, which is adaptive in nature and also Gaussian membership function is used. This technique uses the hybrid optimization which consists of combination of backpropagation and least square method algorithm. Simulation result is shown.