Title :
A systolic architecture for new split algorithms for arbitrary Toeplitz-plus-Hankel matrices
Author :
Fang, Wen-Hsien ; Yagle, Andrew E.
Author_Institution :
Dept. of Electron. Eng., Nat. Taiwan Inst. of Technol., Taipei, Taiwan
fDate :
2/1/1994 12:00:00 AM
Abstract :
Recently, new fast algorithms have been developed for computing the optimal linear least-squares prediction filters for arbitrary Toeplitz-plus-Hankel covariances. The authors propose a systolic architecture that can fully express the inherent concurrency of this highly parallelizable algorithm. The simplification of this array structure for centrosymmetric covariances is also addressed
Keywords :
digital filters; filtering and prediction theory; least squares approximations; matrix algebra; parallel algorithms; systolic arrays; Toeplitz-plus-Hankel matrices; centrosymmetric covariances; concurrency; covariance matrix; fast algorithms; linear least-squares prediction filters; optimal filters; parallel algorithm; split algorithms; systolic architecture; Artificial intelligence; Computer architecture; Concurrent computing; Equalizers; Filters; Iterative algorithms; Signal processing; Signal processing algorithms; Vectors; Very large scale integration;
Journal_Title :
Signal Processing, IEEE Transactions on
