Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices
On-device artificial intelligence has attracted attention globally, and attempts to combine the internet of things and TinyML (machine learning) applications are increasing. Although most edge devices have limited resources, time and energy costs are important when running TinyML applications. In th...
Guardado en:
Autores principales: | , |
---|---|
Formato: | article |
Lenguaje: | EN |
Publicado: |
MDPI AG
2021
|
Materias: | |
Acceso en línea: | https://doaj.org/article/937ae8a4f6e948c0a545db9d1d08abf5 |
Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
id |
oai:doaj.org-article:937ae8a4f6e948c0a545db9d1d08abf5 |
---|---|
record_format |
dspace |
spelling |
oai:doaj.org-article:937ae8a4f6e948c0a545db9d1d08abf52021-11-25T16:43:43ZHardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices10.3390/app1122110732076-3417https://doaj.org/article/937ae8a4f6e948c0a545db9d1d08abf52021-11-01T00:00:00Zhttps://www.mdpi.com/2076-3417/11/22/11073https://doaj.org/toc/2076-3417On-device artificial intelligence has attracted attention globally, and attempts to combine the internet of things and TinyML (machine learning) applications are increasing. Although most edge devices have limited resources, time and energy costs are important when running TinyML applications. In this paper, we propose a structure in which the part that preprocesses externally input data in the TinyML application is distributed to the hardware. These processes are performed using software in the microcontroller unit of an edge device. Furthermore, resistor–transistor logic, which perform not only windowing using the Hann function, but also acquire audio raw data, is added to the inter-integrated circuit sound module that collects audio data in the voice-recognition application. As a result of the experiment, the windowing function was excluded from the TinyML application of the embedded board. When the length of the hardware-implemented Hann window is 80 and the quantization degree is <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mn>2</mn><mrow><mo>−</mo><mn>5</mn></mrow></msup></semantics></math></inline-formula>, the exclusion causes a decrease in the execution time of the front-end function and energy consumption by 8.06% and 3.27%, respectively.Jisu KwonDaejin ParkMDPI AGarticleTinyMLembedded systemfield programmable gate array (FPGA)microcontroller unit (MCU)voice recognitioninter-IC sound (I2S)TechnologyTEngineering (General). Civil engineering (General)TA1-2040Biology (General)QH301-705.5PhysicsQC1-999ChemistryQD1-999ENApplied Sciences, Vol 11, Iss 11073, p 11073 (2021) |
institution |
DOAJ |
collection |
DOAJ |
language |
EN |
topic |
TinyML embedded system field programmable gate array (FPGA) microcontroller unit (MCU) voice recognition inter-IC sound (I2S) Technology T Engineering (General). Civil engineering (General) TA1-2040 Biology (General) QH301-705.5 Physics QC1-999 Chemistry QD1-999 |
spellingShingle |
TinyML embedded system field programmable gate array (FPGA) microcontroller unit (MCU) voice recognition inter-IC sound (I2S) Technology T Engineering (General). Civil engineering (General) TA1-2040 Biology (General) QH301-705.5 Physics QC1-999 Chemistry QD1-999 Jisu Kwon Daejin Park Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices |
description |
On-device artificial intelligence has attracted attention globally, and attempts to combine the internet of things and TinyML (machine learning) applications are increasing. Although most edge devices have limited resources, time and energy costs are important when running TinyML applications. In this paper, we propose a structure in which the part that preprocesses externally input data in the TinyML application is distributed to the hardware. These processes are performed using software in the microcontroller unit of an edge device. Furthermore, resistor–transistor logic, which perform not only windowing using the Hann function, but also acquire audio raw data, is added to the inter-integrated circuit sound module that collects audio data in the voice-recognition application. As a result of the experiment, the windowing function was excluded from the TinyML application of the embedded board. When the length of the hardware-implemented Hann window is 80 and the quantization degree is <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mn>2</mn><mrow><mo>−</mo><mn>5</mn></mrow></msup></semantics></math></inline-formula>, the exclusion causes a decrease in the execution time of the front-end function and energy consumption by 8.06% and 3.27%, respectively. |
format |
article |
author |
Jisu Kwon Daejin Park |
author_facet |
Jisu Kwon Daejin Park |
author_sort |
Jisu Kwon |
title |
Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices |
title_short |
Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices |
title_full |
Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices |
title_fullStr |
Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices |
title_full_unstemmed |
Hardware/Software Co-Design for TinyML Voice-Recognition Application on Resource Frugal Edge Devices |
title_sort |
hardware/software co-design for tinyml voice-recognition application on resource frugal edge devices |
publisher |
MDPI AG |
publishDate |
2021 |
url |
https://doaj.org/article/937ae8a4f6e948c0a545db9d1d08abf5 |
work_keys_str_mv |
AT jisukwon hardwaresoftwarecodesignfortinymlvoicerecognitionapplicationonresourcefrugaledgedevices AT daejinpark hardwaresoftwarecodesignfortinymlvoicerecognitionapplicationonresourcefrugaledgedevices |
_version_ |
1718413032319614976 |