Evaluating Multi-Sensor Placement and Neural Network Architectures for Physical Activity Level Classification

Bo Cui; Xiaowen Song; Tabak Monique; Bert-Jan van Beijnum; Ying Wang

doi:10.48550/arXiv.2502.14434

Evaluating Multi-Sensor Placement and Neural Network Architectures for Physical Activity Level Classification

Bo Cui, Xiaowen Song, Tabak Monique, Bert-Jan van Beijnum, Ying Wang

Research output: Working paper › Preprint › Academic

14 Downloads (Pure)

Abstract

Accurate physical activity level (PAL) classification could be beneficial for osteoarthritis (OA) management. This study examines the impact of sensor placement and deep learning models on AL classification using the Metabolic Equivalent of Task values. The results show that the addition of anankle sensor (WA) significantly improves the classification of intensity activities compared to wrist-only configuration(53% to 86.2%). The CNN-LSTM model achieves the highest accuracy (95.09%). Statistical analysis confirms multi-sensor setups outperform single-sensor configurations (p < 0.05). The WA configuration offers a balance between usability and accuracy, making it a cost-effective solution for AL monitoring, particularly in OA management.

Original language	Undefined
Publisher	ArXiv.org
Number of pages	5
DOIs	https://doi.org/10.48550/arXiv.2502.14434
Publication status	Published - 20 Feb 2025

Keywords

eess.SP

Access to Document

10.48550/arXiv.2502.14434Licence: CC BY-NC-ND

2502.14434v1Final published version, 277 KBLicence: CC BY-NC-ND

Cite this

@techreport{f0d7082e1dcf4100a883b707f4b78d13,

title = "Evaluating Multi-Sensor Placement and Neural Network Architectures for Physical Activity Level Classification",

abstract = " Accurate physical activity level (PAL) classification could be beneficial for osteoarthritis (OA) management. This study examines the impact of sensor placement and deep learning models on AL classification using the Metabolic Equivalent of Task values. The results show that the addition of anankle sensor (WA) significantly improves the classification of intensity activities compared to wrist-only configuration(53\% to 86.2\%). The CNN-LSTM model achieves the highest accuracy (95.09\%). Statistical analysis confirms multi-sensor setups outperform single-sensor configurations (p < 0.05). The WA configuration offers a balance between usability and accuracy, making it a cost-effective solution for AL monitoring, particularly in OA management. ",

keywords = "eess.SP",

author = "Bo Cui and Xiaowen Song and Tabak Monique and \{van Beijnum\}, Bert-Jan and Ying Wang",

year = "2025",

month = feb,

day = "20",

doi = "10.48550/arXiv.2502.14434",

language = "Undefined",

publisher = "ArXiv.org",

type = "WorkingPaper",

institution = "ArXiv.org",

}

TY - UNPB

T1 - Evaluating Multi-Sensor Placement and Neural Network Architectures for Physical Activity Level Classification

AU - Cui, Bo

AU - Song, Xiaowen

AU - Monique, Tabak

AU - van Beijnum, Bert-Jan

AU - Wang, Ying

PY - 2025/2/20

Y1 - 2025/2/20

N2 - Accurate physical activity level (PAL) classification could be beneficial for osteoarthritis (OA) management. This study examines the impact of sensor placement and deep learning models on AL classification using the Metabolic Equivalent of Task values. The results show that the addition of anankle sensor (WA) significantly improves the classification of intensity activities compared to wrist-only configuration(53% to 86.2%). The CNN-LSTM model achieves the highest accuracy (95.09%). Statistical analysis confirms multi-sensor setups outperform single-sensor configurations (p < 0.05). The WA configuration offers a balance between usability and accuracy, making it a cost-effective solution for AL monitoring, particularly in OA management.

AB - Accurate physical activity level (PAL) classification could be beneficial for osteoarthritis (OA) management. This study examines the impact of sensor placement and deep learning models on AL classification using the Metabolic Equivalent of Task values. The results show that the addition of anankle sensor (WA) significantly improves the classification of intensity activities compared to wrist-only configuration(53% to 86.2%). The CNN-LSTM model achieves the highest accuracy (95.09%). Statistical analysis confirms multi-sensor setups outperform single-sensor configurations (p < 0.05). The WA configuration offers a balance between usability and accuracy, making it a cost-effective solution for AL monitoring, particularly in OA management.

KW - eess.SP

U2 - 10.48550/arXiv.2502.14434

DO - 10.48550/arXiv.2502.14434

M3 - Preprint

BT - Evaluating Multi-Sensor Placement and Neural Network Architectures for Physical Activity Level Classification

PB - ArXiv.org

ER -