DS/CDMA信号盲检测与分离技术研究

【作者】 刘飞；

【导师】 郭伟；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2006， 硕士

【摘要】 直接序列扩频通信具有低截获概率,抗干扰能力强以及实现码分多址等优点,因而在抗干扰通信及民用移动通信中得到广泛的应用。相应地,扩频通信的电子对抗和民用无线电资源监管成为非常重要的研究课题,其中对噪声和干扰中直扩码分多址(DS/CDMA)信号的盲检测和分离是首要任务。直扩通信信号检测的研究已发展出许多方法,如能量法、相关法、倒谱法、周期谱法及高阶谱法等。这些方法的提出源于对DS信号的检测和估计,而很少研究将这些方法应用到对DS/CDMA信号的检测和估计。同时,目前有很多针对盲源分离的方法,但对DS条件下的多用户分离,实际上并没有给出一个理想的解决方案。而不管是神经元网络法,ICA法还是最大熵法一般情况下都难以在低信噪比情况下正确分离出混合的DS/CDMA信号。为解决DS/CDMA信号共性参数的估计,本文首先将三种检测方法:平方律法、周期谱法和倒谱法,应用到DS/CDMA信号的检测与参数估计中。其中,平方律法可以提取信号的载波频率,算法简单有效,但不能识别信号是否为扩频信号;周期谱法可以估计信号的PN码速率,但算法复杂,运算量大;倒谱法运算速度快,能估计PN码时间周期,结合周期谱法估计出的PN码速率,可进一步估计出PN码的长度。然后,针对异步的CDMA信号,采用延迟相乘的方法检测用户数目,同时估计出各用户信号的幅值和相位;该方法的缺点是各用户信号必须有一定的相差,同时用户数不能太多。另外,本文还尝试利用CDMA信号的相位重复特性,解决各用户PN序列的分离问题;但这种方法只能在有一定先验知识的条件下,对包含较少用户的DS/CDMA信号有较好效果,能够简单快速的实现用户PN序列的分离与估计。更多还原

【Abstract】 Direct sequence spread spectrum (DS) signals have been widely used for secure communication and mobile communication known as Code Division Multiple Access (CDMA) system. Because DS signals have many advantages such as anti-jamming capability, low probability of interception, multiple access capability and so on. Accordingly, the interception and radio monitoring for spread spectrum communication have been of great research interests. And the chief task is to detect and separate the DS/CDMA signals hidden in the noise and interference.Several methods of sequence spread spectrum signals direction have been developed, such as energy detection, correlation detection, cepstrum detection, cyclic spectral, higher order cumulant detection, etc. These methods developed from the research of detection and estimation for DS signals. But applying these methods to detect the DS/CDMA signals have not been done and analyses have not been made. At the same time, there are many methods for blind source separation, but there is not an ideal solution to multiuser separation in DS condition. Regardless using Nerve Net, ICA or ME method, it is difficult to separate the mixed DS/CDMA signal in low SNR condition.At first, to estimating common parameters, in this dissertation three detection methods-square detection, spectrum correlation detection and cepstrum detection are applied to detect the DS/CDMA signal. These three methods have different advantages and disadvantages respectively. Every method can estimate a different parameter of the signal with good performance. The square detection can picks up the carrier frequencies, which is simple and has great performance. But this method can’t recognize the signal as spread spectrum signal. The spectral correlation detection method can estimates the chip rate of signal’s PN sequence. The shortcoming of spectral detection is its complexity and in need of long time to compute. A simple and powerful approach is cepstrum detection method, which can estimate PN sequence’s time period. Using this parameter and the PN chip rate estimated by spectral correlation method, we can get another parameter, length of the PN sequence. And then, we utilize Delay and更多还原