Title :
Pipelined Architecture for a Radix-2 Fast Walsh–Hadamard–Fourier Transform Algorithm
Author :
Jinqi Liu ; Qianjian Xing ; Xiaobo Yin ; Xiubin Mao ; Feng Yu
Author_Institution :
Dept. of Instrum. Sci., Zhejiang Univ., Hangzhou, China
Abstract :
This brief proposes an efficient radix-2 single-path delay commutator (SDC) pipelined architecture to implement the fast Walsh-Hadamard-Fourier transform (FWFT) algorithm. The proposed architecture includes (log2 N - 1) SDC stages, which are implemented by merged half-butterfly. The merged half-butterfly is proposed to achieve 100% hardware utilization and minimum buffer usage by sharing common merged half-butterflies in the time-multiplexed approach. Compared with the conventional pipelined radix-2 FFT+Walsh-Hadamard Transform (WHT) designs, the proposed architecture reduces the number of buffers by 50% and of adders by 25%. The required number of complex multipliers is decreased to 0.5 log2 N - 0.5, which is roughly the minimum number. Moreover, the proposed architecture can be applied to FFT/WHT/sequence-ordered complex Hadamard transform (SCHT).
Keywords :
Hadamard transforms; buffer circuits; fast Fourier transforms; hypercube networks; multiplying circuits; pipeline processing; FFT-WHT-sequence-ordered complex Hadamard transform; buffer usage; common merged half-butterflies; complex multipliers; hardware utilization; pipelined architecture; radix-2 fast Walsh-Hadamard-Fourier transform; radix-2 single-path delay commutator; time-multiplexed approach; Adders; Computer architecture; Delays; Discrete Fourier transforms; Signal processing algorithms; Fast Walsh???Hadamard???Fourier transform (FWFT); fast Walsh-Hadamard-Fourier transform(FWFT); merged half-butterfly; single-path delay commutator; single-path delay commutator (SDC);
Journal_Title :
Circuits and Systems II: Express Briefs, IEEE Transactions on
DOI :
10.1109/TCSII.2015.2456371