No icon

A Noise Resistant Correlation Method for Period Detection of Noisy Signals

A Noise Resistant Correlation Method for Period Detection of Noisy Signals


This paper develops a novel method called noise resistant correlation (NRC) method for detecting the hidden period from the contaminated (noisy) signals with strong white Gaussian noise. A novel correlation function is proposed based on a newly constructed periodic signal and the contaminated signal to effectively detect the target hidden period. In contrast with the conventional autocorrelation analysis (AUTOC) method, this method demonstrates excellent performance, especially when facing strong noise. Fault diagnoses of rolling element bearings and gears are presented as application examples and the performance of the proposed method is compared with that of the AUTOC method.



The time-frequency domain analysis method uses hybrid period detectors which incorporate features of both the time-domain and frequency-domain approaches.

The time-frequency analysis method can provide local information of source signal within a certain time range or frequency range. Thus, this approach is mainly used to handle signals whose frequencies vary with time and hence is not suitable for our cases.

The time-domain analysis method operates directly on the waveform of the signal to estimate the period and thus provides a potential way to detect the hidden period from strong white noise.


A wide variety of solutions have been proposed to address the above problems. proposed a novel difference function called the average magnitude autocorrelation function (AMDF) to improve the period detection power.

The AMDF is a variation of the autocorrelation function in which a difference signal is formed between the delayed and contaminated signals, and the absolute magnitude is taken at each delay value.

Proposed a new period extraction method which used an autocorrelation function weighted by the inverse of the AMDF to improve the accuracy of the period extraction effect.

Comment As:

Comment (0)