冲击噪声背景下归一化广义旁瓣相消器

Normalized-generalized Sidelobe Canceller in Impulsive Noise

针对冲击噪声背景下常规波束形成算法性能下降的问题,该文提出一种归一化广义旁瓣相消器(N-GSC)算法,该算法适用于任意未知统计特性的代数拖尾冲击噪声环境。算法通过对输入信号进行无穷范数归一化,使信号的二阶统计量存在且有界,再进行维纳滤波,提高了波束形成在冲击噪声背景下的性能。进行了4种冲击噪声背景下的仿真实验。仿真结果表明,与传统的GSC算法和基于分数低阶矩的GSC算法相比,N-GSC算法计算简单,无需噪声特征指数的先验信息或估计,适用于任意分布的冲击噪声环境,具有更强的干扰抑制能力。

To solve the performance degradation of a beamformer in impulsive noises,a new normalized-generalized sidelobe canceller（N-GSC）algorithm is presented for the heavy-tailed impulsive noises of arbitrary unknown statistics.The second-order statistical entity of input signal is made exist and finite by infinity-norm normalizing the input signal,and the input signal is filtered by wiener filter to improve the performance of the beamformer amid heavy-tailed impulsive noise of unknown statistics.The simulation results of four impulsive noises show that,compared with the GSC and the fractional lower order moments based GSC algorithm,the N-GSC algorithm does not need any prior information and estimation of the impulsive noise characteristic exponents,is easy for calculation and suitable for wider heavy-tailed impulsive noises,and can offer better interference-rejection.