A new noise reduction method for chaotic signals: improved local-projection algorithm

The paper addresses the problem of noise reduction for chaotic signals. Until now, there has been no efficient way to direct the noise filtering process. This is why some noise reduction methods always face the trade-off between distorting the original chaotic signal or holding relatively more noise in the filtered signal. On the basis of extensive analysis of local correlation of the chaotic dynamic system, a new method for analyzing a signal's deterministic degree is given. A parameter D, which is very sensitive to noise, is introduced and used to show how deterministic the system is. With the help of parameter D, we improve our former local-projection noise reduction method. Numerical results show the efficiency of the new method. Since this method does not require prior knowledge of the underlying dynamic system, it is a kind of practical method satisfying engineering demands.