Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge

The escalating challenges of managing vast sensor-generated data, particularly in audio applications, necessitate innovative solutions. Current systems face significant computational and storage demands, especially in real-time applications like gunshot detection systems (GSDS), and the proliferatio...

Full description

Saved in:
Bibliographic Details
Main Authors: Sanggeon Yun, Ryozo Masukawa, Hanning Chen, Sungheon Jeong, Wenjun Huang, Arghavan Rezvani, Minhyoung Na, Yoshiki Yamaguchi, Mohsen Imani
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10896650/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850022870801973248
author Sanggeon Yun
Ryozo Masukawa
Hanning Chen
Sungheon Jeong
Wenjun Huang
Arghavan Rezvani
Minhyoung Na
Yoshiki Yamaguchi
Mohsen Imani
author_facet Sanggeon Yun
Ryozo Masukawa
Hanning Chen
Sungheon Jeong
Wenjun Huang
Arghavan Rezvani
Minhyoung Na
Yoshiki Yamaguchi
Mohsen Imani
author_sort Sanggeon Yun
collection DOAJ
description The escalating challenges of managing vast sensor-generated data, particularly in audio applications, necessitate innovative solutions. Current systems face significant computational and storage demands, especially in real-time applications like gunshot detection systems (GSDS), and the proliferation of edge sensors exacerbates these issues. This paper proposes a groundbreaking approach with a near-sensor model tailored for intelligent audio-sensing frameworks. Utilizing a Fast Fourier Transform (FFT) module, convolutional neural network (CNN) layers, and HyperDimensional Computing (HDC), our model excels in low-energy, rapid inference, and online learning. It is highly adaptable for efficient ASIC design implementation, offering superior energy efficiency compared to conventional embedded CPUs or GPUs, and is compatible with the trend of shrinking microphone sensor sizes. Comprehensive evaluations at both software and hardware levels underscore the model’s efficacy. Software assessments through detailed ROC curve analysis revealed a delicate balance between energy conservation and quality loss, achieving up to 82.1% energy savings with only 1.39% quality loss. Hardware evaluations highlight the model’s commendable energy efficiency when implemented via ASIC design, especially with the Google Edge TPU, showcasing its superiority over prevalent embedded CPUs and GPUs.
format Article
id doaj-art-3beb311db2814c42bb6aae13b83f8946
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-3beb311db2814c42bb6aae13b83f89462025-08-20T03:01:32ZengIEEEIEEE Access2169-35362025-01-0113439474395510.1109/ACCESS.2025.354323210896650Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme EdgeSanggeon Yun0https://orcid.org/0000-0002-0488-9666Ryozo Masukawa1https://orcid.org/0009-0009-8988-1388Hanning Chen2https://orcid.org/0000-0003-1956-5135Sungheon Jeong3https://orcid.org/0000-0003-3540-7065Wenjun Huang4https://orcid.org/0009-0005-7416-3587Arghavan Rezvani5https://orcid.org/0000-0002-1012-5785Minhyoung Na6Yoshiki Yamaguchi7https://orcid.org/0009-0001-5485-0399Mohsen Imani8https://orcid.org/0000-0002-5761-0622Department of Computer Science, University of California at Irvine, Irvine, CA, USADepartment of Computer Science, University of California at Irvine, Irvine, CA, USADepartment of Computer Science, University of California at Irvine, Irvine, CA, USADepartment of Computer Science, University of California at Irvine, Irvine, CA, USADepartment of Computer Science, University of California at Irvine, Irvine, CA, USADepartment of Computer Science, University of California at Irvine, Irvine, CA, USADepartment of Computer Science, Kookmin University, Seoul, Republic of KoreaDepartment of Electronic Information Systems, Shibaura Institute of Technology, Saitama, JapanDepartment of Computer Science, University of California at Irvine, Irvine, CA, USAThe escalating challenges of managing vast sensor-generated data, particularly in audio applications, necessitate innovative solutions. Current systems face significant computational and storage demands, especially in real-time applications like gunshot detection systems (GSDS), and the proliferation of edge sensors exacerbates these issues. This paper proposes a groundbreaking approach with a near-sensor model tailored for intelligent audio-sensing frameworks. Utilizing a Fast Fourier Transform (FFT) module, convolutional neural network (CNN) layers, and HyperDimensional Computing (HDC), our model excels in low-energy, rapid inference, and online learning. It is highly adaptable for efficient ASIC design implementation, offering superior energy efficiency compared to conventional embedded CPUs or GPUs, and is compatible with the trend of shrinking microphone sensor sizes. Comprehensive evaluations at both software and hardware levels underscore the model’s efficacy. Software assessments through detailed ROC curve analysis revealed a delicate balance between energy conservation and quality loss, achieving up to 82.1% energy savings with only 1.39% quality loss. Hardware evaluations highlight the model’s commendable energy efficiency when implemented via ASIC design, especially with the Google Edge TPU, showcasing its superiority over prevalent embedded CPUs and GPUs.https://ieeexplore.ieee.org/document/10896650/Audio sensingnear-sensor intelligent sensinghyperdimensional computingASIC design
spellingShingle Sanggeon Yun
Ryozo Masukawa
Hanning Chen
Sungheon Jeong
Wenjun Huang
Arghavan Rezvani
Minhyoung Na
Yoshiki Yamaguchi
Mohsen Imani
Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge
IEEE Access
Audio sensing
near-sensor intelligent sensing
hyperdimensional computing
ASIC design
title Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge
title_full Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge
title_fullStr Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge
title_full_unstemmed Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge
title_short Hyperdimensional Intelligent Sensing for Efficient Real-Time Audio Processing on Extreme Edge
title_sort hyperdimensional intelligent sensing for efficient real time audio processing on extreme edge
topic Audio sensing
near-sensor intelligent sensing
hyperdimensional computing
ASIC design
url https://ieeexplore.ieee.org/document/10896650/
work_keys_str_mv AT sanggeonyun hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT ryozomasukawa hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT hanningchen hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT sungheonjeong hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT wenjunhuang hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT arghavanrezvani hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT minhyoungna hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT yoshikiyamaguchi hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge
AT mohsenimani hyperdimensionalintelligentsensingforefficientrealtimeaudioprocessingonextremeedge