Font Size: a A A

Fault Diagnosis Of Avionic Devices Based On Information Fusion Technology

Posted on:2015-01-28Degree:MasterType:Thesis
Country:ChinaCandidate:X J HuFull Text:PDF
GTID:2272330431993620Subject:Microelectronics and Solid State Electronics
Abstract/Summary:PDF Full Text Request
Avionics equipment is an important part of modern fighter, with the rapid development of science and technology, avionics device is becoming more and more complex, traditional methods of fault diagnosis have been unable to meet the needs of reality. Information fusion technology can integrated redundancy and complementary information from multiple information sources and obtain comprehensive and accurate information of the diagnosed object, to get a more accurate judgment. This text is combined the fault diagnosis of avionic device with Information fusion, a method and model which the fault diagnosis system based on fuzzy neural network and D-S evidence theory is built.In the following for the problem that neural network and fuzzy theory are insufficient and complementary, the fuzzy neural network structure is adopted as part of feature-level information fusion. With the increase of the network input, the network structure becomes more and more complex and due to the fault which the network generalization ability is poor. The structure which the clustering algorithm is introduced becomes optimization. Use an example to get the fault diagnosis, and the results show that this structure has the advantage of fast learning speed、high diagnostic accuracy by comparison with BP neural network and RBF neural network.In the following for the problem that the fault diagnosis has the uncertainty phenomenon, the part of decision-level information fusion adopts D-S evidence theory. As for the question that result from D-S evidence theory cannot obtain the real answer when there are conflicts among evidence, the method is called two-stage modification of evidence source is used. The first stage is based on the characteristics of the output the model accuracy in the process of basic probability assignment of access; the second stage is based on the close degree among evidences as the phenomenon of high conflict among evidences. Finally, use an example to demonstrate the effectiveness of this method. In order to further improve the network generalization ability, the system turn a large-scale fuzzy neural network into a number of small-scale fuzzy neural network according to the parameter space, D-S evidence theory can be integrated redundancy of multiple sub-fuzzy neural network output, and fuzzy neural network can solve the fault that basic probability assignment obtain difficult, the use of the two together can play their respective advantages and strengths. Finally for example in aviation radio, to verify the method of fuzzy neural network and D-S evidence has high accuracy and precision.
Keywords/Search Tags:Fault Diagnosis avionics equipment, Information Fusion, Fuzzy theory, Neural Network, D-S Evidence Theory
PDF Full Text Request
Related items