BLIND EXTRACTION ALGORITHM OF THE HARMONIC SIGNAL BASED ON THE STEADY-STATE POINT CAPTURE IN LORENZ ENERGY ACCUMULATION AREA

摘要

The chaotic signal has the characteristics of the low-frequency,wide spectrum and high energy,which brings great difficulty for extracting the harmonic signal in its energy accumulation area.We take independent component analysis as a framework in this paper.According to the three-dimensional chaotic dynamics system of Lorenz,we propose the steady-state point capture based on the Lyapunov exponent,and then separate blind source for Lorenz energy accumulation area of multi-frequency harmonic signal and Gaussian signal.In different separation algorithms,we respectively compare the improvement of separation performance from steady-state point capture algorithm,and also analyzes the influence of the amplitude,frequency interval of harmonic signal and noise variance on the extraction performance.Computer simulation verified that blind extraction algorithm based on the steady-state point capture can accurately extract the harmonic signal in the case of spectrum aliasing.It provides a new idea for the harmonic signal extraction from chaotic background.