基于奇异值分解的多卫星信号盲检测

Blind Detection of Multi-satellite Signals Based on SVD

针对非合作低信噪比环境下的卫星通信信号检测问题,在信号子空间维数估计的基础上,提出了一种基于奇异值分解的多卫星信号盲检测方法。该方法充分利用奇异值与特征值之间的关系,设计检测统计量将多个信号能量集中起来进行考虑,以适应更低的信噪比,并从理论上对检测性能进行了推导分析。仿真结果表明,该方法简单高效,针对不同的卫星信号,在虚警概率小于1%、信噪比为-11 dB时,盲检测概率均可达90%以上;同时能够在低信噪比环境下适应多信号环境,且其计算量相对特征值方法减少了一个数量级,更适合应用在星载设备上。

This paper introduces a method for blind detection of multi-satellite signals based on singular value decomposition（SVD） with low signal-to-noise ratio（SNR） in non-cooperative systems, by estimating the dimen- sion of signal subspace. The proposed method establishes a test statistic by making full use of the relationship between singular value and eigenvalue, which can gather multiple signal energy. This method can be applied to lower SNR, and the detection performance is analyzed theoretically. Simulation results for common satellite sig- nal indicate that the method is efficient and can adapt to multi-satellite signals under lower SNR, when false alarm probability is set to less than 1%, the blind detection probability reaches above 90% at SNR= - 11 dB. And the computation complexity is reduced tenfold, which can be used in satellite-berne equipment.