A Time-scale Algorithm Based on Wavelet Entropy

Abstract

A wavelet-entropy-based time-scale algorithm is proposed on the basis of wavelet analysis and the fundamental principle of entropy for the analysis of the international atomic time ATI. Through a multi-scale decomposition of atomic clock signals and in the light of energy partition, wavelet entropy is obtained by synthesizing the wavelet coefficients of all scales, and then the atomic clock is weighted by the wavelet entropy. The algorithm is tested with some measured data obtained in a certain laboratory, and the result shows that this method can synthesize the wave frequency characteristics of the atomic clock signals on different scales, and can obviously increase the degree of stability of the time scale. © 2007.