[논문]Energy-Efficient Approximate Speech Signal Processing for Wearable Devices

Park, Taejoon; Shin, Kyoosik; Kim, Nam Sung

doi:10.4218/etrij.17.0116.0462

Abstract ▼ AI-Helper

As wearable devices are powered by batteries, they need to consume as little energy as possible. To address this challenge, in this article, we propose a synergistic technique for energy-efficient approximate speech signal processing (ASSP) for wearable devices. More specifically, to enable the effi...

As wearable devices are powered by batteries, they need to consume as little energy as possible. To address this challenge, in this article, we propose a synergistic technique for energy-efficient approximate speech signal processing (ASSP) for wearable devices. More specifically, to enable the efficient trade-off between energy consumption and sound quality, we synergistically integrate an approximate multiplier and a successive approximate register analog-to-digital converter using our enhanced conversion algorithm. The proposed ASSP technique provides ~40% lower energy consumption with ~5% higher sound quality than a traditional one that optimizes only the bit width of SSP.

주제어

AI 본문요약
AI-Helper

* AI 자동 식별 결과로 적합하지 않은 문장이 있을 수 있으니, 이용에 유의하시기 바랍니다.

가설 설정

In this article, we observe that: a) a successive approximate register (SAR) ADC, which is one of the most commonly used ADCs in wearable devices, can provide key information for approximate multiplications (that is, the leading-one bit (LOB) position of converted digital values), and thus further simplify an approximate multiplier. Secondly, b) audio signals often have very small amplitudes (that is, the most significant bit(MSB) of converted digital values is zero), while the SAR ADC always turns on/off the largest capacitor to determine the MSB, wasting energy.

제안 방법

If the compared input voltage value is larger (less) than 1/2 V_REF, the MSB is set to one (remains zero). After the determination of the MSB, the binary search algorithm is applied to determine the next MSB by comparing V_IN with 3/4 (1/4) of V_REF. This involves switching the second largest capacitor.
In this paper, we exploited AFT of SSP algorithms to minimize the energy consumption of wearable devices. Considering that the ADC and multiplier are the most energyhungry components in advanced SSP, we synergistically integrated a SAR ADC with our enhanced algorithm and an approximate multiplier to provide a strong trade-off between the energy consumption and perceptive sound quality.
1. Overview of method used to augment a speech signal processing (SSP)circuit, for example, adaptive feedback cancellation(AFC), using the proposed approximate SSP (ASSP) technique, which synergistically integrates an approximate multiplier and a SAR ADC.
The feedback effects were generated according to a realistic feedback model [4], and were mixed with the speech samples. The speech samples with the feedback effects were applied to our proposed SAR ADC on-the-fly, and produced digital samples according to our proposed algorithm. They were then fed to the AFC constructed using the aforementioned approximate multiplier.
To evaluate the energy consumption of SAR ADC, we implemented both conventional and enhanced algorithms in Matlab based on an energy-dissipation model for capacitor switching obtained from [9], and we simulated them using the same speech samples obtained from OSR [5]. Table 2 summarizes the energy consumption of both SAR ADC algorithms under four (n, m) pairs: (14, 7), (14, 9), (16, 8), and (16, 10) in both normal and loud speech environments.

성능/효과

Compared with a conventional SAR ADC algorithm, our proposed algorithm consumed 40%–52% (30.2%–46%) less energy per conversion for n = 14 (n = 16), and had a negligible negative impact on the conversion accuracy.

참고문헌 (9)

R. Hegde and N.R. Shanbhag, "Energy-Efficient Signal Processing Via Algorithmic Noise-Tolerance," IEEE/ACM Int. Symp. Low Power Electron. Design, San Diego, CA, USA, Aug. 16-17, 1999, pp. 30-35.
S. Narayanamoorthy et al., "Energy-Efficient Approximate Multiplication for Digital Signal Processing and Classification Applications," IEEE Trans. Very Large Scale Integr. Syst., vol. 23, no. 6, June 2015, pp. 1180-1184.

상세보기
F.M. Yaul and A.P. Chandrakasan, "A 10 b 0.6 nW SAR ADC with Data-Dependent Energy Savings Using LSB-First Successive Approximation," IEEE Int. Solid-State Circuits Conf. Digest Tech. Papers, San Francisco, CA, USA, Feb. 9-13, 2014, pp. 198-199.
A. Spriet et al., "Adaptive Feedback Cancellation in Hearing Aids," J. Franklin Inst., vol. 343, no. 6, Sept. 2006, pp. 545-573.

상세보기
Open Speech Repository, American English. http://www.voiptroubleshooter.com/open_speech/american.html
IEEE Subcommittee on Subjective Measures, "IEEE Recommended Practice for Speech Quality Measurements," IEEE Trans. Audio Electroacoustics, vol. 17, no. 3, Sept. 1969, pp. 225-246.

상세보기
Y. Hu and P.C. Loizou, "Evaluation of Objective Quality Measures for Speech Enhancement," IEEE Trans. Audio, Speech Language Process., vol. 16, no. 1, Jan. 2008, pp. 229-238.

상세보기
Psytechnics, PESQ: an Introduction. http://www.sageinst.com/downloads/960B/wp_pesq_introduction.pdf
P.J.A. Harpe et al., "A 26 W 8 bit 10 MS/s Asynchronous SAR ADC for Low Energy Radios," IEEE J. Solid-State Circuits, vol. 46, no. 7, July 2011, pp. 1585-1595.

상세보기

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 논문명, 저널/프로시딩명, 저자 , 발행년, 권, 호, 시작페이지, 끝페이지, 발행기관 관리번호, 논문명, 대등논문명, 저자 , 저널/프로시딩명, 발행기관, 발행년, 발행언어, 권, 호, 시작페이지, 끝페이지, ISBN, ISSN, 주제분야, 키워드, 초록(한글), 초록(영문), 저자(소속기관)
저장형식	Text(ASCII format) Excel format RefWorks Direct Export RIS format (for Reference Manager, ProCite, EndNote), Scholar's Aids, Mendeley
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증

[국내논문] Energy-Efficient Approximate Speech Signal Processing for Wearable Devices 원문보기

Abstract ▼ AI-Helper

주제어

AI 본문요약
AI-Helper

가설 설정

제안 방법

성능/효과

참고문헌 (9)

이 논문을 인용한 문헌

저자의 다른 논문 :

관련 콘텐츠

원문 보기

원문 URL 링크

오픈액세스(OA) 유형

이 논문과 함께 이용한 콘텐츠

AI-Helper ※ AI-Helper는 오픈소스 모델을 사용합니다.

선택된 텍스트

연합인증

[국내논문] Energy-Efficient Approximate Speech Signal Processing for Wearable Devices 원문보기

Abstract ▼ AI-Helper

주제어

AI 본문요약 엑셀 다운로드 AI-Helper

가설 설정

제안 방법

성능/효과

참고문헌 (9)

이 논문을 인용한 문헌

저자의 다른 논문 :

박태준 (1)

관련 콘텐츠

원문 보기

원문 URL 링크

오픈액세스(OA) 유형

이 논문과 함께 이용한 콘텐츠

AI-Helper ※ AI-Helper는 오픈소스 모델을 사용합니다.

선택된 텍스트

AI 본문요약
AI-Helper