Contribution
Lossless source coding using the Hilbert-Huang transform for artificial multi-channel data
* Presenting author
Abstract:
A method is described for the lossless compression of signals of non-natural origin that occur in industrial environments, for example. The Hilbert-Huang transform is used to decompose the signal into harmonics with exponential amplitudes. The signal residues are encoded based on entropy and signals from other channels are correlated. This means that basic features like frequencies, start times and decay rates are available for compression as well as for subsequent analyses using neural networks, for example. This reduces the overall system-wide computing effort from data acquisition to feature recognition. The compression rates for various data sets are presented.