针对微弱直扩信号扩频码的盲估计和信息码的盲解扩问题,本文提出了一种能同时分离直扩信号扩频码和信息码的非线性盲自适应恒模算法,达到了对直扩信号盲处理。本文首先提出了直扩信号的盲分离问题,然后详细分析推导了盲自适应随机梯度恒模算法,最后将该盲自适应随机梯度恒模算法应用到了对微弱直扩信号的盲分离中,并从理论上阐明了可以用该算法来实现直扩信号的盲分离。所提出的算法完全不同于以往的基于矩阵分解(奇异值分解、特征分解等)的伪码盲估计方法,它的存储开销量和计算量都比较小,可以实现对较长伪码构造的直扩信号的处理,而且它的计算速度较快,在某种程度上解决了传统的基于矩阵分解的方法在直扩信号的实时处理及实现上的困难。理论分析和数值结果都表明了所提方法能较好地工作在较低的输入信噪比条件下。
In this paper, we propose an approach that can blindly separate the PN sequence and message sequence in weak DS- SS signals. This paper proposes the problem of DS-SS signal blind separation firstly, then analyzes the blind adaptive method of stochastic gradient constant modulus algorithm (CMA) in detail, and applies the CMA to the question of DS-SS signal blind separation in the end. This method is totally different from the existed method, which based on decomposition of matrix. It possesses a lot of advantages. Theoretical analysis and experimental results show that the approach can work well for weak DS-SS signals.