Efficient multi-moment non-linear filtering: Experimental study for low power chaotic signals

Abstract

This paper presents experimental results related to the multi-moment non-linear filtering of chaotic signals. For simplicity of implementation only the particular case of two moments filtering (2MM) is developed here. The 2MM filtering approach is specially suited for estimation of extremely weak chaotic signals immersed on an accompanying noise signal (Additive White Gaussian Noise, AWGN) or together with some other type of information signal. The performance of the 2MM technique is rather efficient in the sense that very small values of the Normalized Mean Square Error (NMSE) can be achieved for Chaos/Noise or Chaos/signal ratios below 0 dB. This experimental study allows establishing clearly the trends of the multi-moment approach.