C-RAN预处理单元设计与实现

利用FPGA（现场可编程门阵列）实现基于LTE（Long Term Evolution，长期演进）协议C-RAN（云无线接入网络）体系架构中的前端预处理单元来加速CRAN系统的处理速度。软件层运行在基带信号处理单元池中，并且和FPGA前端预处理单元协同完成基带信号处理的整个过程。其中FPGA前端预处理单元集成了丰富的接口资源，包括PCIE、10 Gb/s以太网口、CPRI（通用公共无线接口）接口，将传统的移动通信和高速数据处理单元有效连接起来。预处理单元通过完成基带处理中的关键算法来减轻服务器处理的压力。与此同时在预处理单元内完成循环前缀的去除和有效子载波数据的筛选，降低了系统的IO吞吐量。预处理单元已经完成了仿真与验证。

C-RAN front-end processing design and implement

This paper presents C-RAN（Cloud Radio Access Network） front-end processing based on LTE by altera FPGA. Motivation of the research work is to accelerate the CRAN. The software stack is running on BBU（BaseBand Unit） pool servers, together with the FPGA front-end processing board, emulates the whole baseband signal processing flow. The FPGA board integrates rich communication interface, including PCI Express, 10 Gb/s Ethernet, CPRI（Common Public Radio Interface） optical connection etc., which enable it to bridge traditional telecom domain and IT data center domain. As an accelerator, the FPGA implements many key algorithm units required by baseband signal processing, which increases the system computing power greatly. Most importantly, with the innovative data pre-processing algorithms（Cyclic Prefix removal） the front-end processing can address the critical system bottleneck on host memory bandwidth and IO throughput. Front-end processing has been verified and tested.