中文摘要：

针对宽带功率放大器的强记忆效应特性，提出一种功放建模和数字预失真方法——PGsc模型。利用广义记忆多项式（GMP）、特定交叉项（sc7）及记忆时刻信号交叉项（CIMT）3个基函数来构造功放行为模型及数字预失真器，并搭建实际测试平台对模型的精度及线性化效果进行验证。测试结果表明，与PMEC方法相比，PGSC方法建模时的归一化均方误差减少了2．1dB，数字预失真时输出信号的三阶邻信道功率比降低了4．94／2．03、dB；与GMP方法相比，PGSC方法仅利用73％的系数即可得到更高的模型精度和更好的线性化效果。

英文摘要：

A new modeling and Digital Pre-distortion（DPD） method named as Parallel GMP-SCT-CIMT（PGSC） is proposed because of the strong memory effect characteristics of broadband power amplifier. The new behavioral model of power amplifier and pre-distorter are constructed of Generalized Memory Polynomial（GMP）, Specific Cross Terms（SCT） and Cross Items between Memory Times（CIMT）. A test platform is built to verify the accuracy and linear effect of the method. Experimental results show that compared with the PMEC method, the PGSC method shows an average Normalized Mean Square Error（NMSE） improvement of 2.1 dB in behavioral modeling, and reduces the third-order ACPR of output signal by 4.94/2.03 dB in DPD application. Compared with the GMP method, the PGSC method achieves higher model accuracy and better linear effect just using approximately 73% of coefficients.