论文信息 - A G-transform based systolic array for Least Squares problems

A G-transform based systolic array for Least Squares problems

A new type of systolic array for the solution of Least Squares problems, based on the G transform, is presented. The G-transformation matrices are permuted Hessenberg matrices and their use for the solution of LS problems based on a systolic array implementation offers some advantages as square-root free implementation and reduced number of multiplications.

Nikos I. Passas | Evangelos Zervas | Athanassia Alonistioti

[1] H. T. Kung,et al. Matrix Triangularization By Systolic Arrays , 1982, Optics & Photonics.

[2] J. G. McWhirter,et al. A novel algorithm and architecture for adaptive digital beamforming , 1986 .

[3] J. G. McWhirter,et al. Recursive Least-Squares Minimization Using A Systolic Array , 1983, Optics & Photonics.

[4] Fuyun Ling,et al. A recursive modified Gram-Schmidt algorithm for least- squares estimation , 1986, IEEE Trans. Acoust. Speech Signal Process..