基于盲源分离的雷达信号欺骗干扰抑制

针对脉冲压缩信号的频谱弥散(SMSP)干扰,通过分析其时频特性,提出了一种基于盲分离的抑制算法。首先,推导了基于独立分量分析的代价函数,由于代价函数中信号概率密度函数不能精确估计,利用信号高阶互累积量的性质,重新构造了一种新的代价函数;其次,利用基于自然梯度的方法对新代价函数进行优化,从而得到基于峭度的盲分离算法;最后,通过仿真证明,对比Fast ICA算法,文中算法具有更快的收敛速度和更好的分离性能,在保证分离精度的前提下,可成功将目标回波信号分离,达到SMSP干扰抑制的目的。

Radar Signal Deception Jamming Suppressing Based on Blind Source Separation

Through analysis time and frequency domains character of the smeared spectrum(SMSP)jamming which aimedat pulse compression signal, a suppressing algorithm was proposed based on blind source separation. First, the contrastfunction was deduced based on independent component analysis(ICA), because the probability density function couldnotbe computed accurately, a new contrast function was constructed by using the character high order cumulant. Second, thenew contract function was optimized by using natural gradient, and a new blind source separation algorithm was proposed.Last, the experimental simulation proved that the algorithm had faster convergence character and better separation perfor.mance than FastICA algorithm, and could separate the target echo from mixed signals with good precision, the purpose tosuppress SMSP jamming was achieved.