基于小波分解和Duffing振子的变尺度微弱信号检测

Weak Signal Detection Based on Scale Transformation of Duffing Oscillator and Wavelet Decomposition

摘要:

传统的时频分析方法在对周期性微弱信号进行检测时,提取的信息具有信噪比不高的缺点,从而影响了检测效果,为此,利用Duffing振子混沌系统对噪声的强免疫力的特征,提出了一种基于小波分解和混沌阵子的混合微弱信号检测方法;首先,采用小波变换对信号进行分解,通过小波变换的平滑作用实现对含噪微弱信号的离散处理,并设计了一种根据阈值来确定分解层数的方法,然后将降噪后的重构信号作为Duffing阵子的周期驱动力并入混沌系统,采用混沌Duffing阵子阵列实现在强噪声背景下的微弱信号检测,并提出了一种临界状态策动力幅值和初始相位的自适应确定方法；在Matlab7仿真环境下进行实验,结果表明:文中方法能有效地对湮没在强噪声下的微弱信息进行检测,具有信号检测信噪比高,重构信号频率较其它方法更接近于真实频率,具有较强的可行性。

Abstract:

Aiming at the traditional time frequency analysis method detecting the cyclic weak signal detection has the defects such as the low SNR (signal to noise ratio) and affecting the detection effect, therefore, using the property of strong immunity of Duffing oscillator chaos system, a compound weak signal detection method based on wavelet decomposition and chaos oscillator. Firstly, the signal is decomposed by wavelet transform and operated discretely based on the smooth effect of wavelet transform, and a method for obtaining the numbers of decomposition layers is designed, then the de-noise re-constructed signal is input to the chaos system as the cycle driving force as the doffing oscillator, and the weak signal detection can be realized by chaos doffing, and a adaptive method based on critical state driving force and initial phase. The experiment is operated in the Matlab7 environment and the results shows that the method can implement the weak signal detection, has the high SNR, the frequency of re-constructed signal is nearly the same as the true signal frequency, has the strong feasibility.