复包络数字滤波器的最大似然辨识建模

Maximum Likelihood Identification and Modeling of Digital Co mplex Envelope Filter

针对无线通讯系统和雷达系统的数字仿真中大量存在的复包络数字滤波器建模问题 ,提出了一种基于变量误差模型 (Errors- in- Variables Model,简记为 EV模型 )的频域最大似然参数估计方法。首先将实验测量值的中心频率移到零频 ,得到复包络滤波器的频率响应 ,然后使用最小二乘法计算复包络滤波器系数的初始估计值 ,最后使用数值迭代法对最大似然估计的代价函数进行优化 ,得到复包络滤波器系数的最佳估计。实验结果以及理论分析表明 ,与现有的复包络数字滤波器转换方法相比 ,最大似然法建模的过程更加直接 。

Complex envelope digital filter is widely used i n computer simulation of modern wireless communication systems and radar systems .A frequency domain maximum likelihood identification method based on errors in variables model(EV model)for the complex envelope digital filter is presented .At first,the carrier frequency of the experimental data are removed And then t he initial estimation of the complex coefficients of the digital filter are calc ulated by the use of least square method.An iterative procedure is used to minim ize the cost function of the maximum likelihood estimation,and results in optim ized complex coefficients of complex envelope filter.Experiments and analyses sh ow that compared with the traditional methods,the new method is more convenient, gives more accurate results.