Bay-Fi: Approximate Bayesian main Frequency Identification for extremely undersampled signals

Bay-Fi (Bayesian Frequency Identification) is a novel method for the direct identification of the dominant modal frequency from extremely under-sampled vibration signals. The approach employs an approximate Bayesian framework to optimize a curve-fitting algorithm applied to sub-Nyquist, randomly acquired measurements, enabling accurate frequency estimates without signal reconstruction or prior hyperparameter tuning.