基于信息融合技术的航空电子设备故障诊断研究  

Fault Diagnosis of Avionic Devices Based on Information Fusion Technology

作　　者：闫涛[1] 赵文俊[1] 胡秀洁[2] 宋家友[2] 

机构地区：[1]空军第一航空学院航空电子工程系,河南信阳464000 [2]郑州大学信息工程学院,郑州450001

出　　处：《电子科技大学学报》2015年第3期392-396,共5页Journal of University of Electronic Science and Technology of China

基　　金：军内计划项目(KJ2012255)

摘　　要：提出一种基于模糊神经网络局部诊断和D-S融合决策的航空电子设备故障诊断方案。首先将设备故障特征信息进行有效组合,构建多个子模糊神经网络完成故障的局部诊断,以获得彼此独立的证据;然后应用Dempster-Shafe证据理论对各个证据进行决策融合,最后根据构建的诊断框架特点用局部诊断输出值及其正确率来获取基本概率赋值。该方案应用于某型号机载电台的仿真结果表明,经过信息融合多级处理后,诊断结论的可信度明显增加,故障识别的正确率得到有效提高。A fault diagnosis scheme for airborne avionics is proposed based on local fault detecting with fuzzy neutral network and decision fusion with Dempster-Shafer evidence theory. Firstly, the characteristic malfunction information of equipment is effectively recombined, and fuzzy neural sub-networks are constructed to achieve independent evidences, with which the diagnosis conclusions as decision fusion results are then drawn by using D-S evidence theory. Lastly, the basic probability values are computed according to the local diagnosis outputs and their credibility. Simulation results indicate that the diagnosis credibility can be obviously increased and the accuracy can also be effectively improved when the scheme is applied to the fault diagnosis of an airborne radio.

关 键 词：航空电子 证据理论 故障诊断 模糊神经网络 信息融合 

分 类 号：TN85[电子电信—信息与通信工程]