시계열 신호의 혼돈분석 기법 소개: 해양 수중소음 신호를 중심으로

Abstract

Ambient noise as a background noise in the ocean has been well known for its the various and irregular signal characteristics. Generally, these signals are treated as noise and they are analyzed through stochastical level if they don't include definite sinusoidal signals. This study is to see how ocean ambient noise can be analyzed by the chaotic analysis technique. The chaotic analysis is carried out with underwater ambient noise obtained in areas near the Korean Peninsula. The calculated physical parameters of time series signal are as follows: histogram, self-correlation coefficient, delay time, frequency spectrum, sonogram, return map, embedding dimension, correlation dimension, Lyapunov exponent, etc. We investigate the chaotic pattern of noises from these parameters. From the embedding dimensions of underwater noises, the assesment of underwater noise by chaotic analysis shows similar results if they don't include a definite sinusoidal signal. However, the values of Lyapunov exponent (divergence exponent) are smaller than that of random noise signal. As a result, we confirm the possibility of classification of underwater noise using Lyapunov analysis.