2019 IEEE International Solid- State Circuits Conference - (ISSCC) 2019
DOI: 10.1109/isscc.2019.8662540
|View full text |Cite
|
Sign up to set email alerts
|

17.2 A 142nW Voice and Acoustic Activity Detection Chip for mm-Scale Sensor Nodes Using Time-Interleaved Mixer-Based Frequency Scanning

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
20
0

Year Published

2019
2019
2023
2023

Publication Types

Select...
5
3

Relationship

0
8

Authors

Journals

citations
Cited by 47 publications
(20 citation statements)
references
References 3 publications
0
20
0
Order By: Relevance
“…Current low power VAD implementations reach less than 1µW , but at a performance cost: e.g. in [31] babble noise at 5dB scores 84% and 72% on hit rate and correct rejections whereas similar conditions ("CAFE" of [26]) on SNN h1 reach 97%, 84%; other power comparisons can be found in table 1 of [31].…”
Section: Resultsmentioning
confidence: 99%
“…Current low power VAD implementations reach less than 1µW , but at a performance cost: e.g. in [31] babble noise at 5dB scores 84% and 72% on hit rate and correct rejections whereas similar conditions ("CAFE" of [26]) on SNN h1 reach 97%, 84%; other power comparisons can be found in table 1 of [31].…”
Section: Resultsmentioning
confidence: 99%
“…Besides, the indispensable Fast Fourier Transform (FFT) in conventional digital signal processing realization costs most processing time [13]. Some work [16]- [18] proposed the analog feature extraction method to avoid A/D conversion. Nevertheless, in their work, the simple features extracted in the analog domain are only suitable for simple tasks such as voice activity detection.…”
Section: (A)mentioning
confidence: 99%
“…Used features are direct FFT bins, frequency band powers, and mel frequency cepstral coefficients (MFCC). Mixed-signal systems employ an analog feature extractor (AFE) which extracts frequency band power features [1], [2], [8], [9]. Classifiers used in aforementioned systems range from simple decision trees (DT) to artificial neural networks.…”
Section: Introductionmentioning
confidence: 99%