宽带通信中数字预失真技术研究与验证

发布时间：2018-03-27 17:36

本文选题：宽带　切入点：数字预失真　出处：《电子科技大学》2017年硕士论文

【摘要】：功率放大器作为通信系统的核心单元,其性能一直是研发人员重要的关注点。在通信信号带宽持续增长的背景下,功放的非线性失真问题将愈加严重。目前,数字预失真技术已经成为补偿功放非线性失真的主流技术。本文主要围绕宽带通信中数字预失真技术展开,包括功率放大器的行为模型提取和数字预失真模型修剪两个方面。对于宽带通信中功率放大器的行为模型提取,本文搭建了一个单输入单输出射频器件特征提取平台。整个平台使用了16位1Gbps采样率的DAC和14位250Mbps采样率的ADC,可工作在400MHz-4GHz频率范围内,实现基带带宽达125MHz。文中使用ADI公司的FMCOMMS1射频评估板和Xilinx公司的高性能FPGA开发板VC707实现了一个宽带高效的零中频收发单元,并对其发射和接收通路中存在的本振泄露和I/Q不平衡问题,以及行为模型建立时存在的延时对齐问题进行了深入的分析,并实现了相应的补偿。宽带通信中功放建模需要更多的非线性项,这会导致计算复杂度的增加。本文使用OMP(Orthogonal Matching Pursuit,正交追踪基)算法和VS-SAMP(Variable Step-size Sparsity Adaptive Matching Pursuit,变步长稀疏度自适应匹配追踪)算法对功放行为模型进行修剪。通过仿真可以看到,OMP和VS-SAMP均可在保证功放建模精度的情况下有效地减少模型系数个数,即使在功放模型的系数个数(稀疏度)未知的情况下,VS-SAMP辨识算法仍能通过变步长的方式,使用尽可能少的系数满足功放建模精度的要求。最后,本文使用VS-SAMP算法实现了数字预失真器,实验验证了其对混合类功放失真的改善效果。实验结果表明,在使用20M和40M LTE信号进行测试时,功率放大器输出信号的ACLR(Adjacent Channel Leakage Ratio,邻信道泄露功率比)可改善16dB,达到-45dBc以下,实现了预期的效果。相比于传统的LS辨识算法,VS-SAMP算法使用了更少的系数。
[Abstract]:Power amplifier as the core unit of communication system, its performance has been an important concern of researchers. In the context of the continuous growth of communication signal bandwidth, the nonlinear distortion of power amplifier will become more serious. Digital predistortion technology has become the mainstream technology to compensate nonlinear distortion of power amplifier. It includes two aspects: the behavior model extraction of power amplifier and the pruning of digital predistortion model. In this paper, a single input and single output RF device feature extraction platform is built. The whole platform uses 16-bit 1Gbps sampling rate DAC and 14-bit 250Mbps sampling rate ADC.The platform can work in the 400MHz-4GHz frequency range. The baseband bandwidth of 125MHz is achieved. In this paper, a wideband and efficient zero-if transceiver unit is implemented by using ADI's FMCOMMS1 radio frequency evaluation board and Xilinx's high performance FPGA development board VC707. The problems of local oscillation leakage and I / Q imbalance in the transmission and reception channels, and the problem of delay alignment when the behavior model is established, are analyzed in detail. And the corresponding compensation is realized. In wideband communication, more nonlinear items are needed for power amplifier modeling. This will increase the computational complexity. This paper uses OMP(Orthogonal Matching pursuit (orthogonal tracking basis) algorithm and VS-SAMP(Variable Step-size Sparsity Adaptive Matching pursuit (variable step size sparse adaptive matching tracking) algorithm to prune the performance model of power amplifier. Both OMP and VS-SAMP can effectively reduce the number of model coefficients under the condition of ensuring the modeling accuracy of power amplifier. Even when the number of coefficients (sparsity) of the power amplifier model is unknown, the VS-SAMP identification algorithm can use as few coefficients as possible to meet the requirements of the modeling accuracy of the amplifier by changing the step size. Finally, In this paper, the digital predistorter is realized by using VS-SAMP algorithm, and the effect of improving the distortion of hybrid power amplifier is verified by experiments. The experimental results show that, when using 20m and 40M LTE signals, the digital predistorter is tested. The ACLR(Adjacent Channel Leakage Ratio (adjacent channel leakage power ratio) of the output signal of the power amplifier can be improved by 16dBs, below -45dBc, and the expected effect is achieved. Compared with the traditional LS identification algorithm, the VS-SAMP algorithm uses less coefficients.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN722.75;TN929.5

【参考文献】