Novel hybrid approach for fault diagnosis in 3-DOF flight Simulator based on BP neural network and ant colony algorithm

In the 3-DOF(degree-of-freedom) flight simulator system, the relations between observed information and fault causes are very complicated. Based on the description of the basic principle of the ant colony algorithm, a novel hybrid approach for fault diagnosis in 3-DOF flight simulator is proposed in this paper, which is based on BP(back propagation) neural network and ant colony algorithm. Combining with rough set theory, ant colony algorithm is used to compute the reductions of the decision table. Then, the condition attributes of decision table are regarded as the input nodes of BP neural network and the decision attributes are regarded as the output nodes of BP neural network correspondingly. Experiments demonstrate that the proposed hybrid approach could achieve a fairly good performance, yield good prediction accuracy of the prediction errors.