FPGA Implementation of Matrix Inversion Using QRD-RLS Algorithm
Authors:
- M. KarkootiMKM. Karkooti
- J.R. CavallaroJCJ.R. Cavallaro
- C. DickCDC. Dick
- 25 May 2006
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10586393,p. 1625-1629
Abstract
Conference PaperThis paper presents a novel architecture for matrix inversion by generalizing the QR decomposition-based recursive least square (RLS) algorithm. The use of Squared Givens rotations and a folded systolic array makes this architecture very suitable for FPGA implementation. Input is a 4 Ã 4 matrix of complex, floating point values. The matrix inversion design can achieve throughput of 0.13M updates per second on a state of the art Xilinx Virtex4 FPGA running at 115 MHz. Due to the modular partitioning and interfacing between multiple Boundary and Internal processing units, this architecture is easily extendable for other matrix sizesKeywords
This publication has 10 references indexed in Scilit:
- A road to future broadband wireless access: MIMO-OFDM-Based air interfaceIEEE Communications Magazine, 2005
- Complexity Analysis of MMSE Detector Architectures for MIMO OFDM Systems2005
- Architectures for adaptive weight calculation on ASIC and FPGA2003
- Redundant and on-line CORDIC for unitary transformationsIEEE Transactions on Computers, 1994
- Numerical accuracy and hardware tradeoffs for CORDIC arithmetic for special-purpose processorsIEEE Transactions on Computers, 1993
- A Square Root and Division Free Givens Rotation for Solving Least Squares Problems on Systolic ArraysSIAM Journal on Scientific and Statistical Computing, 1991
- Squared Givens RotationIMA Journal of Numerical Analysis, 1991
- CORDIC arithmetic for an SVD processorJournal of Parallel and Distributed Computing, 1988
- Matrix Triangularization By Systolic Arrays1982
- The CORDIC Trigonometric Computing TechniqueIRE Transactions on Electronic Computers, 1959