Research On Weak Harmonic Signal Detection Algorithm Under Strong Chaotic Background

Current studies show that many engineering problems can be boiled down to the problems of detecting weak harmonic signal from strong chaotic interference.Chaos theory has been widely used in different kinds of fields in society for its universality and rich dynamic characteristics.Thus,the problem of detecting weak harmonic signal in strong chaotic background is a big issue in research.In this paper,Chaos theory and signal detection algorithms are used in the study on weak harmonic signal detection under chaotic background.The main works are summarized as follows:1?Firstly,a method using Duffing chaotic oscillator as the detection system is proposed.It is a find that chaotic oscillators is sensitive to initial condition and the phase plane changed when the periodic driving force is changed.We take the character as the principle to detect weak harmonic signal.The system's phase plane state changes from chaos state to large-scale periodic state if the system parameter has a weak harmonic signal.2?Secondly,three detection methods under homogeneous chaotic background is proposed.The ESN detection method using reference sequence to train the echo state network.Then use the trained network to reconstruct the one-step prediction matrix.Finally,detect the weak harmonic signal by one-step prediction error.However,the ESN detection method is sensitive to white Gaussian noise.In order to solve the problem,we propose an optimization algorithm based on SOCP.Noticing that the chaotic signal's second-order statistics properties are stationary,the SOCP method construct data matrix by using reference sequence.Then an optimal filter is designed to get the optimal weight vector.Finally,we detect the weak harmonic signal by output SINR.Unfortunately,the SOCP method has to solve a convex optimization problem,which has a high computational complexity.To reduce the computational complexity,we proposed an iterative algorithm based on EKF.Compared with traditional EKF method,the proposed one has a lower computational complexity.3?Thirdly,a detection methods under inhomogeneous chaotic background is proposed.The inhomogeneous detection method does not need a reference sequence.First,the inhomogeneous detection method obtains the virtual reference sequence by constructing the Hankel data matrix.Then an M-order optimal FIR filter is designed,which can minimize the output power of chaotic background and pass the weak harmonic signal undistorted.Finally,detection and reconstruction of the weak harmonic signal are achieved from the output SINR.The simulation results prove that the inhomogeneous detection method can detect weak harmonic signal in homogeneous chaotic background and inhomogeneous chaotic background,meanwhile,the inhomogeneous detection method has a good anti-noise performance.