用于非合作信号噪底估计的改进自回归算法

非合作接收中由于信道或接收机模拟器件的影响可能导致接收信号功率谱存在噪声基底不平坦现象，为便于划定检测门限需要先估计噪声基底再对原始功率谱进行修正。传统的非线性自回归滑动滤波算法采用固定阈值，仅适用于窄带单信号的噪底估计。改进算法通过增加调节系数实现对阈值的动态调节，并根据功率谱特征计算初始阈值和调节系数，提高了原算法的适应范围和工程实用性。仿真结果表明，改进算法对不同种类的信号具有普适性，能解决原算法无法适应的宽带多信号噪底估计。

An improved recursive algorithm for noise floor estimation of non-cooperative received signals

Due to the impact of channel or receiver components,the noise floor of received signal spectrum could become uneven. In order to set signal detection threshold conveniently,it is required to estimate noise floor at first,and modify original spectrum afterwards. The classical nonlinear recursive smoothing filter algorithm sets the fixed threshold,so it is only useful for narrow band single signal. The improved algorithm in this paper adjusts the threshold dynamicly by adding adjustable coefficients,and calculates the initial threshold and adjustable coefficient according to original sampling data,which expands the operative range and practicality of the classical algorithm. Through a series of simulation experiments,it is demonstrated that the improved algorithm is useful for various signals. Especially,it is effective to wide band signals,which cannot be processed using the classical algorithm.