Most of the existing techniques for identifying the disturbance signal waveforms of a vehicle electrical system are primarily based on visual inspection. The current paper proposes a wavelet packet decomposition-based technique to perform a feature extraction from the disturbance signal. The disturbance signal of the wiper motor is first decomposed into various frequency sub-band signals, after which a method of calculating the integrated energy and mutation parameters is presented using the wavelet pack coefficient (WPC) of the signal after decomposition. And the feature vector representing the electromagnetic interference (EMI) signal is developed using the extracted parameters. The standard vector library for feature parameters is constructed with the multi-sample averaging method, from which a recognition algorithm with different frequency ranges and categories is proposed. The results are validated using four kinds of EMI signals from the actual vehicle that performed satisfactorily and effectively in the pattern recognition of EMI sources.
Electromagnetic interference pattern recognition for vehicle wiper motor based on wavelet packet decomposition
International Journal of Electric and Hybrid Vehicles (online) ; 4 , 2 ; 186-196
2012
11 Seiten, 6 Bilder, 16 Quellen
Article (Journal)
English
Automotive engineering | 2012
|