应用反馈神经网络分析功放的记忆效应  

Feedback Neural Networks Model in Analyzing Memory Effects in RF Power Amplifiers

作　　者：王鹏[1] 丛志鹏[1] 李平辉[2] 

机构地区：[1]解放军理工大学通信工程学院研究生3队,江苏南京210007 [2]解放军理工大学通信工程学院卫星通信系,江苏南京210007

出　　处：《军事通信技术》2007年第2期50-53,共4页Journal of Military Communications Technology

摘　　要：非线性行为模型在功放的线性化研究中被广泛应用。在大功率功放的使用中,记忆效应也是不能忽略的因素。通过研究Wiener模型和神经网络模型的适用范围和优缺点,提出了用于记忆效应建模的反馈神经网络模型。并且对特定功放应用双音正弦信号和谐波平衡法进行仿真,从而验证了该反馈模型可以很好的描述记忆效应。The nonlinear behavioral models are often used in the linearization of RF power amplifiers. And when establishing the model, the memory effect is not neglected. By studying the characteristic of the parallel Wiener model and Elman model, this paper proposed Elman behavioral model for RF power amplifiers, which exhibit memory effects. The model extraction was performed using two-tone harmonic-balance measurement with different tone frequency spacings and power levels. It is found that by using such a model, the performance has been advanced to a certain extent.

关 键 词：非线性 行为模型 记忆效应 

分 类 号：TN722.7[电子电信—电路与系统]