全双工射频自干扰抑制快速搜索算法研究与实现

发布时间：2018-06-04 01:40

本文选题：全双工 + 射频自干扰抑制　；参考：《电子科技大学》2014年硕士论文

【摘要】：现有无线通信技术为了达到双向通信的目的,一般使用TDD或者FDD来正交划分时频资源,是一种半双工系统。全双工系统的目标是无线设备能够同一时间、同一频点进行通信,使无线通信的频谱利用率提高一倍,然而它的主要阻碍是,来自本节点的发送信号会干扰远端传送过来的有用信号。在全双工模式下,目前已有的抑制自干扰信号技术主要包括天线自干扰抑制、射频自干扰抑制和数字自干扰抑制三种。针对射频自干扰抑制,论文设计并实现了基于快速搜索算法的技术方案,重点讨论了梯度预测的理论依据和优化策略,并利用C语言编程实现了方案的关键技术。在全双工演示样机上对方案的指标需求进行了测试验证,给出了分析结果。首先,研究了现有的全双工自干扰抑制理论和技术手段,尤其是对射频自干扰抑制算法进行了详细讨论,为本课题的研究提供参考和设计思路。根据方案性能指标,提出了使用工程测量的方式实现梯度下降的方案,并利用二次回归模型实现了二维坐标点的预测,给出了完整详细的优化方案。其次,根据硬件结构,实现了射频自干扰抑制快速搜索算法的关键模块。从模块间的接口信息和模块内的处理流程等方面,给出方案C语言实现的详细设计。再次,在详细设计的基础上进行了方案软仿真。利用ZC序列数据,针对不同场景进行了搜索测试,对搜索路径、梯度区域、预测步长和信噪比等参数分别进行了测试,通过软仿真给出了实现过程中各参数取值规律的参考性结论。最后,工程验证了射频自干扰抑制快速搜索算法方案。利用全双工无线通信样机,在不同自干扰信号功率情况下对抑制性能进行了测试,验证了本方案性能。在自干扰信号为?10dBm时,能够达到48dB的抑制度,平均收敛时间小于60s,满足设计需求。本文提出了一种基于梯度预测的射频自干扰抑制方案,研究成果已经得到工程验证。该方法为以后的射频自干扰抑制算法设计提供了理论和实现依据,有助于推动全双工通信模式的进一步发展和应用。
[Abstract]:In order to achieve the purpose of two-way communication, the existing wireless communication technologies generally use TDD or FDD to divide time-frequency resources quadrature, which is a half-duplex system. The goal of a full-duplex system is that wireless devices can communicate at the same time and at the same frequency point, doubling the spectrum efficiency of wireless communications. The transmitted signal from this node interferes with useful signals transmitted from the far end. In full-duplex mode, the existing technologies of suppressing self-interference mainly include antenna self-interference suppression, RF self-interference suppression and digital self-interference suppression. Aiming at RF self-interference suppression, this paper designs and implements the technical scheme based on fast search algorithm, discusses the theoretical basis and optimization strategy of gradient prediction, and realizes the key technology of the scheme by using C language programming. The index requirements of the scheme are tested and verified on the full duplex demo prototype, and the analysis results are given. First of all, the existing full-duplex self-interference suppression theory and technical means are studied, especially the RF self-interference suppression algorithm is discussed in detail, which provides a reference and design ideas for the research of this topic. According to the performance index of the scheme, the scheme of gradient descent by engineering measurement is put forward, and the prediction of two-dimensional coordinate point is realized by using quadratic regression model, and a complete and detailed optimization scheme is given. Secondly, according to the hardware structure, the key module of RF self-interference suppression fast search algorithm is realized. From the interface information between modules and the processing flow in the module, the detailed design of C language implementation is given. Thirdly, based on the detailed design, the soft simulation of the scheme is carried out. Using the ZC sequence data, the search test is carried out for different scenes, and the parameters such as search path, gradient region, prediction step size and signal-to-noise ratio are tested respectively. Through the soft simulation, the paper gives the reference conclusion of the rule of each parameter value in the process of realization. Finally, the project verifies the fast search algorithm of RF self-interference suppression. The performance of the proposed scheme is tested under different self-interference signal power using a full-duplex wireless communication prototype. When the self-interference signal is 10 dBm, the suppression degree of 48dB can be achieved, and the average convergence time is less than 60 s, which meets the design requirements. In this paper, a scheme of RF self interference suppression based on gradient prediction is proposed, and the research results have been verified by engineering. This method provides a theoretical and practical basis for the future design of RF self-interference suppression algorithm, and helps to promote the further development and application of full-duplex communication mode.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.4

【相似文献】