Implementation of an unknown input observer based on Cellular Neural Network
Baya Hajjej, Boumedyen Boussaid, Ahmed Zouinkhi, Mohamed Naceur Abdelkrim · 2017
This paper deals with an improved approach of fault detection based on cellular neural network (CNN). This method is a simplified idea from the original CNN circuit in order to detect faults in real time by hardware implementation. This idea is very interesting for smart sensor design and implementation in embedded systems. In this work, a CNN is simulated as an observer in order to detect faulty sensors. This fault is due to different causes such as, electronic failure, hardware failure or a battery drain. The CNN observer is an analogical circuit which is applied in the system of temperature sensor used in wireless sensor network applications. Two different categories of observers(Luenberger observer and Unknown Input Observer) are given and compared in this work to show the importance of the method. The simulation results by Matlab/Simscape tools shows the effectiveness of the approach.