未来移动通信系统中，高效率和高线性度是功率放大器研究的难点。宽带高效率Doherty功放结合数字预失真技术是目前主要的解决方案之一。随着信号带宽和峰均比的增大，宽带高效率Doherty功放的非线性特性和数字预失真处理都变得十分复杂。因此，本项目重点研究带宽达到100MHz的高效率Doherty功放的非线性特性、行为建模方法和数字预失真技术。采用简单的非线性函数替换高阶项和复杂计算的方法，探索低复杂度、高精度的简化Volterra级数模型；采用模糊神经网络理论和方法，研究新型有记忆的宽带模糊神经网络模型，以及相应的自适应算法；在模型研究的基础上，利用欠采样和查找表技术设计高效、低复杂度和低成本的宽带数字预失真方法和算法。本项目的研究将扩充功放行为建模方法和促进线性化技术的发展。

In future mobile communication systems, high efficiency and linearity are the difficulty points in research of the power amplifiers. Designing high-efficiency wideband Doherty power amplifiers with digital predistortion (DPD) is one of the major solutions at present. With the increase of the signal bandwidth and peak-to-average power ratio (PAPR), the nonlinear characteristics of the high-efficiency wideband Doherty power amplifiers and the processing of DPD have become very complicated. So our project will focus on the nonlinear characteristics, behavioral modeling and DPD of the high-efficiency Doherty power amplifiers with up to 100 MHz signal bandwidth. It will explore the simplified Volterra series model with low complexity and high precision when the high order items and complex calculations are replaced by simple nonlinear functions. The novel fuzzy neural network models with memory effects and the adaptive algorithms will be studied by the theory and method of the fuzzy neural networks. Based on the behavioral models, under-sampling and look-up table will be used to design the wideband DPD method and algorithm with low complexity, low cost and high efficiency. Our work will extend the behavioral modeling methodologies for power amplifiers and promote the development of linearization techniques.