Detection of Weak Pulse Signal under Chaotic Noise based on Fractional Maximum Correlation Entropy Algorithm

Abstract

How to improve the detection accuracy of target weak signal is always the difficulty of signal processing. In this paper, based on fractional maximum correlation entropy algorithm and combined with the local linear model, a method for detecting weak pulse signal in chaotic noise background is proposed. Firstly, for the sensitivity of chaotic signal to initial values and short-term predictability, reconstruct the phase space of the observation signal, establish a local linear model, use the fractional maximum correlation entropy algorithm for parameter estimation, and perform a single-step prediction to obtain the prediction error. Then, in order to accurately detect the submerged weak pulse signal, a threshold is given. Finally, the simulation results show that the proposed model in this paper can effectively detect the weak pulse signal under the background of chaotic noise, and it is suitable for signals of different intensities, and the detection speed and accuracy are much better than other models.

Feature Extraction of Lubricating Oil Debris Signal Based on Segmentation Entropy with an Adaptive Threshold

“…Detection of non-stationary short-time pulse signals is encountered in many areas of engineering applications [ 1 ]. Pulses might be presented in various forms on account of the physical mechanisms [ 2 , 3 , 4 , 5 ].…”

Identification of low-SNR Coulter signals based on strongly coupled Lorenz-like oscillators

“…The above research shows a broad application of chaotic oscillators in weak signal detection, which is still in the empirical exploratory stage [26][27][28], and practical models have rarely been reported thus far. In addition, the detection model is mainly an improved Duffing system, such as Rössler, Lorenz, Chua and their extended models [29][30][31][32], which have rarely been reported thus far. At present, coupled Duffing oscillators have some disadvantages, such as that the system greatly depends on the initial state, has limited anti-noise ability and is insensitive to weak pulse signals with narrow bandwidths.…”