Robust Spectral Peaks Detection in Vibration and Acoustic Signals

This article brings a practical solution to the problem of spectral peak detection in nonuniform spectra. It applies a robust probabilistic approach that fits the histogram of trimmed spectral data with a truncated Gamma distribution. The estimated distribution parameters are used to derive a threshold through a hypothesis test in the presence of peaks. The proposed approach gains its robustness from the formulation of the no-peak distribution, while no knowledge is available about the amount of peaks in spectral data. The authors propose a preprocessing step to cope with a nonuniform spectrum. The proposed methodology is validated on both simulated and experimental vibration and acoustic signals.

Automated summary

This article presents a practical solution to the problem of detecting spectral peaks in nonuniform spectra. A robust probabilistic approach is applied, where the histogram of trimmed spectral data is fitted with a truncated Gamma distribution. The estimated distribution parameters are used to derive a threshold through a hypothesis test. The proposed approach is robust as it formulates the distribution without peaks, regardless of the number of peaks in the spectral data. The authors also suggest a preprocessing step to handle nonuniform spectra, and the methodology is validated using simulated and experimental signals.