ELECTRICAL SOURCES PERFORMANCE IMPROVEMENT OF UNMANNED AIRBORNE VEHILES USING ARTIFICIAL NEURAL NETWORKS BASED SFDIA

Traid Kurdi, Saadi

	ELECTRICAL SOURCES PERFORMANCE IMPROVEMENT OF UNMANNED AIRBORNE VEHILES USING ARTIFICIAL NEURAL NETWORKS BASED SFDIA
AL-TAQANI
Article 1, Volume 28, Issue 1, 2015, Pages 0-0
Author
Saadi Traid Kurdi
Abstract
The occurrence of faults during the flight of unmanned air vehicle UAV is a very critical situation that affects the completion of the mission. It was found that these faults are mainly due to failure in the sender (sensor), it was also found that the rate of failure is high in electrical energy Sensors (ac, dc and battery).This research presents an effective technique to ensure that the electrical power system (.ac, dc, battery) or the sensors are faulting free. This technique using two different approaches. The first approach is Radial Basis Function RBF-NN trained with the Extended Minimal Resource Allocation Network - EMRAN algorithms. The second approach, which is presented in this Paper, is based on Knowledge based Neural Network NN based tool SFDIA (Sensor Failure, Detection, Identification and Accommodation problem). Neural Network ANN based tool SFDIA Sensor Failure, Detection, Identification and Accommodation problem, are used to provide analytical redundancy from which residuals are generated, enabling the detection of failures on sensor signals. Upon detection of failure, the faulty signal is replaced by the neural network based estimate. This technique allows the flight to continue within specified performance limitations. The results achieved from the modeling process showed that the neural network based tool SFDIA is able to show high-resolution results in the behavior of electrical energy Sensors (ac, dc and battery).
Keywords
Unmanned aircraft vehicle UAV; Sensor fault detection SFD; Fault diagnosis; and Aircraft modeling and simulation

Statistics Article View: 144 PDF Download: 72