We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.AR

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Hardware Architecture

Title: A 23 $μ$W Keyword Spotting IC with Ring-Oscillator-Based Time-Domain Feature Extraction

Abstract: This article presents the first keyword spotting (KWS) IC which uses a ring-oscillator-based time-domain processing technique for its analog feature extractor (FEx). Its extensive usage of time-encoding schemes allows the analog audio signal to be processed in a fully time-domain manner except for the voltage-to-time conversion stage of the analog front-end. Benefiting from fundamental building blocks based on digital logic gates, it offers a better technology scalability compared to conventional voltage-domain designs. Fabricated in a 65 nm CMOS process, the prototyped KWS IC occupies 2.03mm$^{2}$ and dissipates 23 $\mu$W power consumption including analog FEx and digital neural network classifier. The 16-channel time-domain FEx achieves 54.89 dB dynamic range for 16 ms frame shift size while consuming 9.3 $\mu$W. The measurement result verifies that the proposed IC performs a 12-class KWS task on the Google Speech Command Dataset (GSCD) with >86% accuracy and 12.4 ms latency.
Comments: 14 pages, 21 figures, 2 tables
Subjects: Hardware Architecture (cs.AR); Sound (cs.SD); Audio and Speech Processing (eess.AS)
DOI: 10.1109/JSSC.2022.3195610
Cite as: arXiv:2208.00693 [cs.AR]
  (or arXiv:2208.00693v1 [cs.AR] for this version)

Submission history

From: Kwantae Kim [view email]
[v1] Mon, 1 Aug 2022 09:04:30 GMT (10013kb,D)

Link back to: arXiv, form interface, contact.