Neural network based architecture for fault detection and isolation in air data systems

❖ A FDI architecture for Air Data Systems measurements based on Artificial Neural Network was presented. ❖ CRISP-DM process was applied to flight data to find the optimal set for the design and test of the Artificial Neural Network. ❖ Different ANNs were designed to ensure a fault tolerant FDI architecture ❖ Experimental results for Indicated Airspeed and Angle of Attack were presented (the architecture was applied also for True Airspeed and Angle of sideslip). ❖ The system was success fully tested on a very light aircraft, whereas the software architecture can be reused for air data systems designed for other classes of aircraft.