考虑记忆交叉项的宽带功放行为建模方法

针对宽带功率放大器的强记忆效应,提出了一种前馈交叉记忆多项式模型。该模型由主信号对消和畸变注入两个环路构成。信号对消环路采用记忆多项式模型对带内动态线性记忆效应进行建模;畸变注入环路采用记忆时刻信号间的交叉记忆项对带外扩展频谱行为进行建模。为克服最小二乘法在模型辨识中带来的数值不稳定问题,采用共轭梯度的方法对模型的参数进行辨识。为消除噪声和测量误差对模型参数辨识带来的影响,采用ADS软件包络仿真的方法获取精确的功放输入输出信号复包络数据,并对该模型进行充分的性能分析和评估。实验结果表明,所提方法对宽带功放的非线性动态特性具有很好的逼近能力,特别是带外频谱扩展特性,较之常用的方法提升了近10 dB。

Wideband power amplifier behavioral modeling including cross memory

Considering the strong memory effects of the broadband power amplifier, a feedforward cross memory polynomial model is proposed. This model consists of two loops. One is used as a signal cancellation loop which adopts memory polynomial to model dynamic linear memory effect in-band. And the other is used as a distortion injection loop, which adopts memory time signals cross item memory to model spectrum spread behavior out-of-band. In order to overcome the numerical instability problem of the least-square method, a conjugate gradient method is used to identify the model coefficients. The ADS envelope simulation method is adopted to obtain accurate complex envelop data of power amplifier input and output signals. This method eliminates the effects of noise and measurement error on the identification of model parameters effectively. The model performance is analyzed and evaluated sufficiently based on the obtained complex envelop data. The experimental results show that the proposed model can achieve good approximation capability for the nonlinear dynamic characteristics of the broadband power amplifier, especially the spectrum spread behavior out-of-band. Approximation precision is improved almost 10 dB compared with the common methods.