水中运动目标动态线谱增强算法研究

An algorithms of underwater moving-target dynamic line spectrum enhancement

通常的基于短时自相关的自适应线谱增强器(SABALSE)主要缺点是:输入信噪比低时,抑制高斯噪声性能差。为了最大限度地克服SABALSE的缺点,我们充分利用高阶累积量抑制高斯噪声的性能和高阶累积量的不同更新算法,提出了基于四阶累积量不同切片的自适应动态线谱增强新算法,并对其原理、结构进行了剖析。用实测鱼雷线谱数据,对鱼雷与水听器处于不同方位时,水听器接收的线谱进行了动态仿真。结果表明:基于四阶累积量非对角切片的自适应动态线谱增强(NDSCBADLSE)算法抑制高斯噪声、增强动态线谱的能力强于基于对角切片的自适应动态线谱增强(DSCBADLSE)算法,且均强于SABALSE算法。因此,本文的算法可用于提高水下探测系统和水下武器系统对微弱信号的检测能力。

Traditional short-term autocorrelation-based adaptive line spectrum enhancer (SABALSE) becomes low in suppressing Gaussian noise when input signal-to-noise ratio becomes low. For greatly overcoming this disadvantage, based on the advantage of higher order cumulants in handling Gaussian noise and their different updating algorithms, the new algorithms of different slices of fourth-order cumulant-based adaptive dynamic line spectrum enhancer are proposed. These algorithms include the algorithm of non-diagonal slice of cumulant-based adaptive dynamic line spectrum enhancer (NDSCBADLSE) and that of diagonal slice of cumulant-based adaptive dynamic line spectrum enhancer (DSCBADLSE). Their conceptual schemes are analyzed in detail. The torpedo-radiated line spectra are enhanced dynamically under condition of different azimuth angles between the torpedo and the hydrophone. The results have shown that the NDSCBADLSE algorithm outperforms the DSCBADLSE algorithm in handling Gaussian noise and enhancing dynamic line spectrum, and that the performance of the NDSCBADLSE and DSCBADLSE algorithm is far better than the SABALSE algorithm. Accordingly, the proposed algorithms may be applied in improving the ability of underwater detection and weapon system to detect underwater target.