混合噪声下正弦参数估计互高阶谱MUSIC方法

采用互高阶累积量 (互四阶累积量 )的一维对角切片为统计量 ,首次证明了互高阶累积量可以有效地抑制非相关噪声和高斯噪声 ,并在建立互高阶累积量的Yule -Walker方程的基础上 ,通过特征分解 ,建立了信号矢量空间与噪声矢量空间 ,首次提出了混合噪声背景下正弦参数估计互高阶谱MUSIC方法。仿真结果说明 ,该方法在几乎不需要有色噪声的先验信息的条件下 ,具有良好的谱估计的分辨率和谱估计的稳定性。与其他方法比较 ,该方法抗干扰性更强 ,其信噪比工作门限低 ,特别适合于工程应用

Music Method of Cross High Order Spectrum for Sinusoidal Parameter Estimation on Hybrid Noise Background

Takes one dimension diagonal slice of cross high order accumulation as statistical element to demonstrate for the first time that cross high order accumulation is able to depress noncorrelative noises and correlative gauss noises. Signal vector space and noise vector space are established through characteristic decomposition, and a novel spectrum analysis method-cross high order spectrum MUSIC method is deduced to estimate sinusoidal parameters on background of hybrid noises. Simulation shows that on case of almost no pre-knowledge about noises, this method exhibits good spectrum estimating differentiability and stability. For low signal to noise ratio level is acceptable, this method is more applicable to practice.