Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (26): 71-75.

Previous Articles     Next Articles

Implementation method of inverse QR decomposition SMI algorithm parallel processing based on FPGA

LIU Qianli   

  1. Military Representative Office of Navy in Wuhan 438 Factory, Wuhan 430060, China
  • Online:2012-09-11 Published:2012-09-21

基于FPGA的逆QR分解SMI算法的并行实现方法

刘千里   

  1. 海军驻武汉四三八厂军事代表室,武汉 430060

Abstract: Based on discussion of the inverse QR decomposition in the SMI adaptive beam-forming algorithm, it studies Systolic array parallel implementation structures of the inverse QR decomposition SMI algorithm, analyzes the PE unit which is used to compose Systolic array to achieve the basic computing modules, gives the inverse QR decomposition SMI algorithm based on Systolic array structure parallel implementation of the FPGA, proposes overall system design and architecture.

Key words: QR decomposition, Sample Matrix Inversion(SMI), adaptive beam-forming, Systolic array, Field Programmable Gate Array(FPGA)

摘要: 在讨论了逆QR分解(逆正交三角分解)SMI(采样矩阵求逆)自适应波束形成算法的基础上,研究了逆QR分解SMI算法的Systolic阵列(脉动阵列)并行实现结构,分析了组成Systolic阵列的各PE(处理单元)单元的基本运算模块的实现,并给出了逆QR分解SMI算法基于Systolic阵列结构的FPGA(现场可编程门阵列)并行实现方法,提出了系统整体的设计与构架。

关键词: 正交三角分解, 采样矩阵求逆, 自适应波束形成, Systolic阵列, 现场可编程门阵列