A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor

This paper presents a method for processing and digitizing the current signal information output from biosensors and a hardware Artificial Intelligence (AI) model design that classifies data using low-power and compact AI algorithms to minimize the high-power consumption and on chip area of the Conv...

Full description

Saved in:
Bibliographic Details
Main Authors: Geon-Hoe Kim, Dong-Gyun Kim, Sung-Jae Lee, Jong-Han Kim, Da-Yeong An, Hyejin Kim, Young-Gun Pu, Heejeong Jasmine Lee, Jun-Eun Park, Kang-Yoon Lee
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11008630/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850123309201489920
author Geon-Hoe Kim
Dong-Gyun Kim
Sung-Jae Lee
Jong-Han Kim
Da-Yeong An
Hyejin Kim
Young-Gun Pu
Heejeong Jasmine Lee
Jun-Eun Park
Kang-Yoon Lee
author_facet Geon-Hoe Kim
Dong-Gyun Kim
Sung-Jae Lee
Jong-Han Kim
Da-Yeong An
Hyejin Kim
Young-Gun Pu
Heejeong Jasmine Lee
Jun-Eun Park
Kang-Yoon Lee
author_sort Geon-Hoe Kim
collection DOAJ
description This paper presents a method for processing and digitizing the current signal information output from biosensors and a hardware Artificial Intelligence (AI) model design that classifies data using low-power and compact AI algorithms to minimize the high-power consumption and on chip area of the Convolutional Neural Network (CNN) models used in biosensors. The dataset was built using a data sampling method that digitizes the current signal values from the biosensor by sampling them 16 times with a 16-bit Analog-to-Digital Converter (ADC), enabling feature extraction in advance. To reduce power consumption and area, a BNN-MLP model without an extraction layer was designed, and to improve accuracy, a dense layer was added to the final layer. This approach enhances accuracy while using binary weights. The BNN-MLP model was designed using TensorFlow for the software model and implemented in hardware at the Register Transfer Level (RTL). To verify the similarity between the hardware-implemented BNN-MLP model and the software BNN-MLP model, classification was performed on a test dataset of 1,000 samples using a Field Programmable Gate Array (FPGA), and the classification accuracy for each class was compared. When the BNN-MLP model is implemented on an FPGA, it utilizes 7,994 Look-Up Tables (LUT) and 9,780 Flip-Flops (FF), occupying fewer resources compared to previous studies. Additionally, it operates at a lower power consumption of 0.157 W, demonstrating the highest power efficiency performance of 407.6 GOPS/W. Finally, the inference time for a single data point was 0.018 ms, which is much faster than in previous studies, confirming its potential for low-power and compact applications in biosensors.
format Article
id doaj-art-c25885e0691a4db59855f1e9a71ee9d1
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-c25885e0691a4db59855f1e9a71ee9d12025-08-20T02:34:38ZengIEEEIEEE Access2169-35362025-01-0113907419075210.1109/ACCESS.2025.357220111008630A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a BiosensorGeon-Hoe Kim0https://orcid.org/0009-0002-3045-505XDong-Gyun Kim1Sung-Jae Lee2https://orcid.org/0009-0008-7771-856XJong-Han Kim3Da-Yeong An4https://orcid.org/0009-0001-4869-2857Hyejin Kim5Young-Gun Pu6Heejeong Jasmine Lee7https://orcid.org/0000-0001-5153-756XJun-Eun Park8https://orcid.org/0000-0001-6345-7903Kang-Yoon Lee9https://orcid.org/0000-0001-9777-6953Department of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South KoreaDepartment of Artificial Intelligence, Sungkyunkwan University, Suwon, South KoreaDepartment of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South KoreaDepartment of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South KoreaSKAIChips, Suwon, South KoreaDepartment of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South KoreaSKAIChips, Suwon, South KoreaSKAIChips, Suwon, South KoreaDepartment of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South KoreaDepartment of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South KoreaThis paper presents a method for processing and digitizing the current signal information output from biosensors and a hardware Artificial Intelligence (AI) model design that classifies data using low-power and compact AI algorithms to minimize the high-power consumption and on chip area of the Convolutional Neural Network (CNN) models used in biosensors. The dataset was built using a data sampling method that digitizes the current signal values from the biosensor by sampling them 16 times with a 16-bit Analog-to-Digital Converter (ADC), enabling feature extraction in advance. To reduce power consumption and area, a BNN-MLP model without an extraction layer was designed, and to improve accuracy, a dense layer was added to the final layer. This approach enhances accuracy while using binary weights. The BNN-MLP model was designed using TensorFlow for the software model and implemented in hardware at the Register Transfer Level (RTL). To verify the similarity between the hardware-implemented BNN-MLP model and the software BNN-MLP model, classification was performed on a test dataset of 1,000 samples using a Field Programmable Gate Array (FPGA), and the classification accuracy for each class was compared. When the BNN-MLP model is implemented on an FPGA, it utilizes 7,994 Look-Up Tables (LUT) and 9,780 Flip-Flops (FF), occupying fewer resources compared to previous studies. Additionally, it operates at a lower power consumption of 0.157 W, demonstrating the highest power efficiency performance of 407.6 GOPS/W. Finally, the inference time for a single data point was 0.018 ms, which is much faster than in previous studies, confirming its potential for low-power and compact applications in biosensors.https://ieeexplore.ieee.org/document/11008630/Binary neural networksartificial intelligencesensor data classificationdata samplingFPGAlow-power
spellingShingle Geon-Hoe Kim
Dong-Gyun Kim
Sung-Jae Lee
Jong-Han Kim
Da-Yeong An
Hyejin Kim
Young-Gun Pu
Heejeong Jasmine Lee
Jun-Eun Park
Kang-Yoon Lee
A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor
IEEE Access
Binary neural networks
artificial intelligence
sensor data classification
data sampling
FPGA
low-power
title A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor
title_full A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor
title_fullStr A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor
title_full_unstemmed A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor
title_short A Low Power Memory-Integrated Hardware BNN-MLP Model on an FPGA for Current Signals in a Biosensor
title_sort low power memory integrated hardware bnn mlp model on an fpga for current signals in a biosensor
topic Binary neural networks
artificial intelligence
sensor data classification
data sampling
FPGA
low-power
url https://ieeexplore.ieee.org/document/11008630/
work_keys_str_mv AT geonhoekim alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT donggyunkim alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT sungjaelee alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT jonghankim alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT dayeongan alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT hyejinkim alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT younggunpu alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT heejeongjasminelee alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT juneunpark alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT kangyoonlee alowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT geonhoekim lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT donggyunkim lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT sungjaelee lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT jonghankim lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT dayeongan lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT hyejinkim lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT younggunpu lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT heejeongjasminelee lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT juneunpark lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor
AT kangyoonlee lowpowermemoryintegratedhardwarebnnmlpmodelonanfpgaforcurrentsignalsinabiosensor