基于最大熵PDF估计的DSSS系统抗窄带干扰技术

Narrowband interference suppression in DSSS system based on maximum entropy PDF estimation

为了抑制直接序列扩频(DSSS)系统中的强窄带干扰,提出了一种新的基于最大熵概率密度函数(PDF)估计的局部最优检测器(LOD)。运用最大熵PDF估计来解析地表达LOD中观测噪声的PDF,并采用一种基于最小二乘法参数初始值设定的非线性Gauss-Newton算法来准确估计最大熵PDF中的拉格朗日系数。所采用的干扰抑制技术减少了LOD中观测噪声PDF估计不准确引入的误差,同时不需要任何训练数据和信号幅值信息。仿真结果表明:基于最大熵PDF估计的LOD干扰抑制技术能够强有力地抑制窄带干扰,复现直扩信号的三角相关特性。当输入干噪比为20dB时,经过此干扰抑制器后接收信号的干噪比改善量相对于线性自适应FIR滤波器和非线性自适应ACM滤波器分别有30 dB和26 dB的改善,同时信噪比损失量分别有3.1 dB和1.6 dB的改善,并且输入干噪比越大,干噪比改善量就越大。

In order to suppress the strong narrowband interference in the direct sequence spread spectrum （DSSS） system, a new locally optimal detector （LOD） based on maximum entropy probability density function （PDF） estimation is proposed. The maximum entropy PDF estimation used to predict and mathematically express the probability that density function of observation noise can be applied into LOD to reduce the error from PDF estimation. Meanwhile, a new nonlinear Gauss-Newton algorithm based on LMS initial parameter design was adopted to estimate the Lagrange weights of maximum entropy PDF. This interference rejection technology didn＇ t need any exercise data and signal amplitude message. The simulation results indicated that the LOD based on maximum entropy PDF estimation can forcefully suppress narrowband interference and represent the triangle relation characteristic of DS signal. When the input JNR is 20 dB, the JNR improvement after this LOD is respectively improved for 30 dB and 26 dB compared with linearly adaptive FIR filter and nonlinearly adaptive ACM filter, meanwhile the SNR is respectively improved for 3.1 dB and 1.6dB, and the JNR improvement is larger for larger input JNR.