High-Accuracy and Area-Efficient Stochastic FIR Digital Filters Based on Hybrid Computation

Shunsuke KOSHITA; Naoya ONIZAWA; Masahide ABE; Takahiro HANYU; Masayuki KAWAMATA

doi:10.1587/transinf.2016LOP0011

High-Accuracy and Area-Efficient Stochastic FIR Digital Filters Based on Hybrid Computation

Shunsuke KOSHITA, Naoya ONIZAWA, Masahide ABE, Takahiro HANYU, Masayuki KAWAMATA

Full Text Views

0

Cite this

Summary :

This paper presents FIR digital filters based on stochastic/binary hybrid computation with reduced hardware complexity and high computational accuracy. Recently, some attempts have been made to apply stochastic computation to realization of digital filters. Such realization methods lead to significant reduction of hardware complexity over the conventional filter realizations based on binary computation. However, the stochastic digital filters suffer from lower computational accuracy than the digital filters based on binary computation because of the random error fluctuations that are generated in stochastic bit streams, stochastic multipliers, and stochastic adders. This becomes a serious problem in the case of FIR filter realizations compared with the IIR counterparts because FIR filters usually require larger number of multiplications and additions than IIR filters. To improve the computational accuracy, this paper presents a stochastic/binary hybrid realization, where multipliers are realized using stochastic computation but adders are realized using binary computation. In addition, a coefficient-scaling technique is proposed to further improve the computational accuracy of stochastic FIR filters. Furthermore, the transposed structure is applied to the FIR filter realization, leading to reduction of hardware complexity. Evaluation results demonstrate that our method achieves at most 40dB improvement in minimum stopband attenuation compared with the conventional pure stochastic design.

Publication: IEICE TRANSACTIONS on Information Vol.E100-D No.8 pp.1592-1602

Publication Date: 2017/08/01

Publicized: 2017/05/22

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2016LOP0011

Type of Manuscript: Special Section PAPER (Special Section on Multiple-Valued Logic and VLSI Computing)

Category: VLSI Architecture

Authors

Shunsuke KOSHITA
  Tohoku University
Naoya ONIZAWA
  Tohoku University
Masahide ABE
  Tohoku University
Takahiro HANYU
  Tohoku University
Masayuki KAWAMATA
  Tohoku University

Keyword

FIR digital filter, stochastic computation, computational accuracy, digital circuit implementation

Cite this

Copy

Shunsuke KOSHITA, Naoya ONIZAWA, Masahide ABE, Takahiro HANYU, Masayuki KAWAMATA, "High-Accuracy and Area-Efficient Stochastic FIR Digital Filters Based on Hybrid Computation" in IEICE TRANSACTIONS on Information, vol. E100-D, no. 8, pp. 1592-1602, August 2017, doi: 10.1587/transinf.2016LOP0011.
Abstract: This paper presents FIR digital filters based on stochastic/binary hybrid computation with reduced hardware complexity and high computational accuracy. Recently, some attempts have been made to apply stochastic computation to realization of digital filters. Such realization methods lead to significant reduction of hardware complexity over the conventional filter realizations based on binary computation. However, the stochastic digital filters suffer from lower computational accuracy than the digital filters based on binary computation because of the random error fluctuations that are generated in stochastic bit streams, stochastic multipliers, and stochastic adders. This becomes a serious problem in the case of FIR filter realizations compared with the IIR counterparts because FIR filters usually require larger number of multiplications and additions than IIR filters. To improve the computational accuracy, this paper presents a stochastic/binary hybrid realization, where multipliers are realized using stochastic computation but adders are realized using binary computation. In addition, a coefficient-scaling technique is proposed to further improve the computational accuracy of stochastic FIR filters. Furthermore, the transposed structure is applied to the FIR filter realization, leading to reduction of hardware complexity. Evaluation results demonstrate that our method achieves at most 40dB improvement in minimum stopband attenuation compared with the conventional pure stochastic design.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2016LOP0011/_p

Copy

@ARTICLE{e100-d_8_1592,
author={Shunsuke KOSHITA, Naoya ONIZAWA, Masahide ABE, Takahiro HANYU, Masayuki KAWAMATA, },
journal={IEICE TRANSACTIONS on Information},
title={High-Accuracy and Area-Efficient Stochastic FIR Digital Filters Based on Hybrid Computation},
year={2017},
volume={E100-D},
number={8},
pages={1592-1602},
abstract={This paper presents FIR digital filters based on stochastic/binary hybrid computation with reduced hardware complexity and high computational accuracy. Recently, some attempts have been made to apply stochastic computation to realization of digital filters. Such realization methods lead to significant reduction of hardware complexity over the conventional filter realizations based on binary computation. However, the stochastic digital filters suffer from lower computational accuracy than the digital filters based on binary computation because of the random error fluctuations that are generated in stochastic bit streams, stochastic multipliers, and stochastic adders. This becomes a serious problem in the case of FIR filter realizations compared with the IIR counterparts because FIR filters usually require larger number of multiplications and additions than IIR filters. To improve the computational accuracy, this paper presents a stochastic/binary hybrid realization, where multipliers are realized using stochastic computation but adders are realized using binary computation. In addition, a coefficient-scaling technique is proposed to further improve the computational accuracy of stochastic FIR filters. Furthermore, the transposed structure is applied to the FIR filter realization, leading to reduction of hardware complexity. Evaluation results demonstrate that our method achieves at most 40dB improvement in minimum stopband attenuation compared with the conventional pure stochastic design.},
keywords={},
doi={10.1587/transinf.2016LOP0011},
ISSN={1745-1361},
month={August},}

Copy

TY - JOUR
TI - High-Accuracy and Area-Efficient Stochastic FIR Digital Filters Based on Hybrid Computation
T2 - IEICE TRANSACTIONS on Information
SP - 1592
EP - 1602
AU - Shunsuke KOSHITA
AU - Naoya ONIZAWA
AU - Masahide ABE
AU - Takahiro HANYU
AU - Masayuki KAWAMATA
PY - 2017
DO - 10.1587/transinf.2016LOP0011
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 8
JA - IEICE TRANSACTIONS on Information
Y1 - August 2017
AB - This paper presents FIR digital filters based on stochastic/binary hybrid computation with reduced hardware complexity and high computational accuracy. Recently, some attempts have been made to apply stochastic computation to realization of digital filters. Such realization methods lead to significant reduction of hardware complexity over the conventional filter realizations based on binary computation. However, the stochastic digital filters suffer from lower computational accuracy than the digital filters based on binary computation because of the random error fluctuations that are generated in stochastic bit streams, stochastic multipliers, and stochastic adders. This becomes a serious problem in the case of FIR filter realizations compared with the IIR counterparts because FIR filters usually require larger number of multiplications and additions than IIR filters. To improve the computational accuracy, this paper presents a stochastic/binary hybrid realization, where multipliers are realized using stochastic computation but adders are realized using binary computation. In addition, a coefficient-scaling technique is proposed to further improve the computational accuracy of stochastic FIR filters. Furthermore, the transposed structure is applied to the FIR filter realization, leading to reduction of hardware complexity. Evaluation results demonstrate that our method achieves at most 40dB improvement in minimum stopband attenuation compared with the conventional pure stochastic design.
ER -