欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2015, Vol. 32 ›› Issue (2): 259-263.DOI: 10.7523/j.issn.2095-6134.2015.02.016

• 信息与电子科学 • 上一篇    下一篇

基于FPGA的高速浮点FFT/IFFT处理器设计与实现

苏斌1,2, 刘畅1, 潘志刚1   

  1. 1. 中国科学院电子学研究所, 北京 100190;
    2. 中国科学院大学, 北京 100049
  • 收稿日期:2014-03-06 修回日期:2014-05-12 发布日期:2015-03-15
  • 通讯作者: 苏斌, subin_iecas@163.com
  • 基金资助:

    国家自然科学基金(61100201)资助

Design and implementation of high-speed floating points FFT processor based on FPGA

SU Bin1,2, LIU Chang1, PAN Zhigang1   

  1. 1. Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China;
    2. University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2014-03-06 Revised:2014-05-12 Published:2015-03-15

摘要:

设计一种基于FPGA的改进的并行FFT/IFFT蝶形运算结构.该结构采用按时间抽选的FFT基-2蝶形算法对IEEE单精度浮点数构成的复数进行8路并行处理.利用Xilinx ISE 13.1软件完成FFT/IFFT处理器的设计, 并在Virtex6硬件平台上进行验证.结果表明, 利用这种8路并行结构设计的FFT/IFFT处理器可在合理利用硬件资源的同时提高运算速度及精度.

关键词: FPGA, 单精度浮点, FFT/IFFT, 基-2蝶形算法, 并行结构

Abstract:

An improved parallel FFT/IFFT butterfly structure is presented. The structure, which is based on the decimation-in-time (DIT) Radix-2 algorithm, contains 8 paths in parallel. IEEE single precision floating complex FFT/IFFT processors are designed by using this parallel structure and achieved with Xilinx ISE 13.1 software. Verifications are carried out with Virtex6 FPGA of Xilinx. The results show that the FFT/IFFT processors can make use of the resources reasonably and enhance both the speed and precision of computation.

Key words: FPGA, single precision floating, FFT/IFFT, Radix-2 algorithm, parallel structure

中图分类号: