Spectral Fault Receptive Fields

Spectral Fault Receptive Fields are intended for analyzing spectral features in signals to support condition monitoring and prognosis.

Spectral Fault Receptive Fields (SFRFs) constitute a computational model of health perception for rotating machinery, inspired by the center-surround receptive fields observed in color-encoding ganglion cells of the primate retina. Opponent receptive fields are defined over structured families of frequency supports associated with characteristic fault frequencies. Receptive-field response functions integrate spectral information, while contrast mappings produce low-dimensional descriptors that aim to capture degradation trends while remaining interpretable by construction. Through flexible parameterizations, SFRFs support the definition of perceptual channels that can be tailored to condition monitoring, diagnosis, and prognosis tasks.

This MATLAB toolbox provides a reference implementation of the SFRF conceptual model and its generalizations, enabling reproducible research and the development of health-monitoring pipelines for rotating machinery, with particular emphasis on rolling-element bearings.