Sensor data correction with neural network incorporating fuzzy logic

This paper details a framework carried out to investigate the possibility of building a combination of neural network and fuzzy logic to address the sensor data correction portion of the "no fault found" problem. The framework is composed of three modules: pre-processing module, fuzzy knowledge based control module, neural network module. The fuzzy knowledge based controller was experimented at the front end of the neural network to isolate drifting sensor values and to correct them before passing to the next stage. Simple simulation results of the online performance of the final fuzzy knowledge based controller-neural network prototype are also presented.