Loading [MathJax]/extensions/tex2jax.js

On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning

Publikation: Beitrag in Fachzeitschrift › Artikel › Forschung › Peer-Review

Autorschaft

Organisationseinheiten

Details

OriginalspracheEnglisch
Seiten (von - bis)404-407
Seitenumfang4
FachzeitschriftCurrent Directions in Biomedical Engineering
Jahrgang10
Ausgabenummer4
PublikationsstatusVeröffentlicht - 1 Dez. 2024

Abstract

Human activity recognition (HAR) is an expanding area of research. Although sensors are becoming more readily available, there is a trend toward minimizing the number of associated sensors for better applicability. Neural networks are often employed for HAR, as they can identify movement patterns within data. To optimize the amount of useful information provided to the network, feature extraction methods are commonly applied. However, these feature methods complicate the data processing pipeline, and thus are less applicable to continuous real-time applications. In this work, we investigate the impact of using quaternions as input features for an LSTM network on HAR, specifically focusing on level walking and stair climbing activities, while also considering the inference time. We demonstrate that combining quaternions and raw IMU data, i. e., acceleration and angular rates significantly enhances classification accuracy without adversely affecting the inference time. By additionally adding an approximate estimate of the vertical position change to the input data, the classification accuracy is further improved to a value of 96.18%.

ASJC Scopus Sachgebiete

Zitieren

On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning. / Kuhlgatz, Timo; Jordine, Marco; Lehmann, Dustin et al.
in: Current Directions in Biomedical Engineering, Jahrgang 10, Nr. 4, 01.12.2024, S. 404-407.

Publikation: Beitrag in Fachzeitschrift › Artikel › Forschung › Peer-Review

Kuhlgatz, T, Jordine, M, Lehmann, D & Seel, T 2024, 'On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning', Current Directions in Biomedical Engineering, Jg. 10, Nr. 4, S. 404-407. https://doi.org/10.1515/cdbme-2024-2099
Kuhlgatz, T., Jordine, M., Lehmann, D., & Seel, T. (2024). On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning. Current Directions in Biomedical Engineering, 10(4), 404-407. https://doi.org/10.1515/cdbme-2024-2099
Kuhlgatz T, Jordine M, Lehmann D, Seel T. On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning. Current Directions in Biomedical Engineering. 2024 Dez 1;10(4):404-407. doi: 10.1515/cdbme-2024-2099
Kuhlgatz, Timo ; Jordine, Marco ; Lehmann, Dustin et al. / On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning. in: Current Directions in Biomedical Engineering. 2024 ; Jahrgang 10, Nr. 4. S. 404-407.
Download
@article{8fc2dd5492394e70a13a8f49465110fe,
title = "On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning",
abstract = "Human activity recognition (HAR) is an expanding area of research. Although sensors are becoming more readily available, there is a trend toward minimizing the number of associated sensors for better applicability. Neural networks are often employed for HAR, as they can identify movement patterns within data. To optimize the amount of useful information provided to the network, feature extraction methods are commonly applied. However, these feature methods complicate the data processing pipeline, and thus are less applicable to continuous real-time applications. In this work, we investigate the impact of using quaternions as input features for an LSTM network on HAR, specifically focusing on level walking and stair climbing activities, while also considering the inference time. We demonstrate that combining quaternions and raw IMU data, i. e., acceleration and angular rates significantly enhances classification accuracy without adversely affecting the inference time. By additionally adding an approximate estimate of the vertical position change to the input data, the classification accuracy is further improved to a value of 96.18%.",
keywords = "Human activity recognition, deep learning, feature assessment, gait classification, quaternions",
author = "Timo Kuhlgatz and Marco Jordine and Dustin Lehmann and Thomas Seel",
note = "{\textcopyright} 2024 by Walter de Gruyter Berlin/Boston",
year = "2024",
month = dec,
day = "1",
doi = "10.1515/cdbme-2024-2099",
language = "English",
volume = "10",
pages = "404--407",
number = "4",

}

Download

TY - JOUR

T1 - On Stair Walk Recognition Using a Single Magnetometer-free IMU and Deep Learning

AU - Kuhlgatz, Timo

AU - Jordine, Marco

AU - Lehmann, Dustin

AU - Seel, Thomas

N1 - © 2024 by Walter de Gruyter Berlin/Boston

PY - 2024/12/1

Y1 - 2024/12/1

N2 - Human activity recognition (HAR) is an expanding area of research. Although sensors are becoming more readily available, there is a trend toward minimizing the number of associated sensors for better applicability. Neural networks are often employed for HAR, as they can identify movement patterns within data. To optimize the amount of useful information provided to the network, feature extraction methods are commonly applied. However, these feature methods complicate the data processing pipeline, and thus are less applicable to continuous real-time applications. In this work, we investigate the impact of using quaternions as input features for an LSTM network on HAR, specifically focusing on level walking and stair climbing activities, while also considering the inference time. We demonstrate that combining quaternions and raw IMU data, i. e., acceleration and angular rates significantly enhances classification accuracy without adversely affecting the inference time. By additionally adding an approximate estimate of the vertical position change to the input data, the classification accuracy is further improved to a value of 96.18%.

AB - Human activity recognition (HAR) is an expanding area of research. Although sensors are becoming more readily available, there is a trend toward minimizing the number of associated sensors for better applicability. Neural networks are often employed for HAR, as they can identify movement patterns within data. To optimize the amount of useful information provided to the network, feature extraction methods are commonly applied. However, these feature methods complicate the data processing pipeline, and thus are less applicable to continuous real-time applications. In this work, we investigate the impact of using quaternions as input features for an LSTM network on HAR, specifically focusing on level walking and stair climbing activities, while also considering the inference time. We demonstrate that combining quaternions and raw IMU data, i. e., acceleration and angular rates significantly enhances classification accuracy without adversely affecting the inference time. By additionally adding an approximate estimate of the vertical position change to the input data, the classification accuracy is further improved to a value of 96.18%.

KW - Human activity recognition

KW - deep learning

KW - feature assessment

KW - gait classification

KW - quaternions

UR - http://www.scopus.com/inward/record.url?scp=85217083855&partnerID=8YFLogxK

U2 - 10.1515/cdbme-2024-2099

DO - 10.1515/cdbme-2024-2099

M3 - Article

VL - 10

SP - 404

EP - 407

JO - Current Directions in Biomedical Engineering

JF - Current Directions in Biomedical Engineering

IS - 4

ER -

Von denselben Autoren