An Analog Circuit Diagnosis Method Based on Adaptive-Kernel ICA
Yongpan Jing 1,
Lifen Yuan 2, and
Yigang He 2
1. Institute of Physics and Information, Hunan Normal University, Changsha, China
2. Institute of Electrical and Automation Engineering, Hefei University of Technology, Hefei, China
2. Institute of Electrical and Automation Engineering, Hefei University of Technology, Hefei, China
Abstract—An analog circuit fault diagnosis method based on an adaptive kernel function is presented. The adaptive kernel function is built by using the linear properties of the kernel functions. There are three steps for analog circuit diagnosis: preprocessing, feature extraction and faults classifying. The data is de-noised and centralized by preprocessing step firstly. Then a feature extraction method based on AMK-ICA technology is presented to lower the dimension of the data. Finally, the support vector machine is used to identify the fault mode of the circuit under test. Experimental study shows that the method can effectively improve the samples training time, test time and identification precision compared with the main reference cited in this paper.
Index Terms—analog circuit, fault diagnosis, independent component analysis, support vector machine
Cite: Yongpan Jing, Lifen Yuan, and Yigang He, "An Analog Circuit Diagnosis Method Based on Adaptive-Kernel ICA," International Journal of Electronics and Electrical Engineering, Vol. 4, No. 2, pp. 116-120, April 2016. doi: 10.18178/ijeee.4.2.116-120
Cite: Yongpan Jing, Lifen Yuan, and Yigang He, "An Analog Circuit Diagnosis Method Based on Adaptive-Kernel ICA," International Journal of Electronics and Electrical Engineering, Vol. 4, No. 2, pp. 116-120, April 2016. doi: 10.18178/ijeee.4.2.116-120
Array