High-speed Cordic based parallel weight extraction for QRD-RLS adaptive filtering

In this paper, a very high-speed Cordic based parallel weight extraction algorithm for QRD-RLS adaptive filtering is developed using the matrix lookahead technique. The algorithm is based on inverse updating and is pipelined at fine-grain level. The proposed architecture can be operated at arbitrarily high sample rates, and consists of only Givens rotations which can be mapped onto a Jacobi specific dataflow processor. It requires a complexity of O(Mp/sup 2/) Givens rotations per sample period, where p is the size of the input data samples, and M is the pipelining level.

[1]  Ed F. Deprettere,et al.  A fully pipelined RLS-based array for channel equalization , 1996, J. VLSI Signal Process..

[2]  Keshab K. Parhi,et al.  Pipelined Cordic based QRD-RLS adaptive filtering using matrix lookahead , 1997, 1997 IEEE Workshop on Signal Processing Systems. SiPS 97 Design and Implementation formerly VLSI Signal Processing.

[3]  Keshab K. Parhi,et al.  Pipeline interleaving and parallelism in recursive digital filters. I. Pipelining using scattered look-ahead and decomposition , 1989, IEEE Trans. Acoust. Speech Signal Process..

[4]  Ed F. Deprettere,et al.  A strategy for determining a Jacobi specific dataflow processor , 1997, Proceedings IEEE International Conference on Application-Specific Systems, Architectures and Processors.

[5]  Keshab K. Parhi,et al.  Pipelined RLS adaptive filtering using scaled tangent rotations (STAR) , 1996, IEEE Trans. Signal Process..

[6]  Charles E. Leiserson,et al.  Optimizing Synchronous Circuitry by Retiming (Preliminary Version) , 1983 .

[7]  Keshab K. Parhi,et al.  Pipeline interleaving and parallelism in recursive digital filters. II. Pipelined incremental block filtering , 1989, IEEE Trans. Acoust. Speech Signal Process..

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