A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition
Sensor-based motion recognition integrates the emerging area of wearable sensors with novel machine learning techniques to make sense of low-level sensor data and provide rich contextual information in a real-life application. Although Human Activity Recognition (HAR) problem has been drawing the at...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2017-03-01
|
Series: | Sensors |
Subjects: | |
Online Access: | http://www.mdpi.com/1424-8220/17/3/529 |
_version_ | 1798043701042216960 |
---|---|
author | Majid Janidarmian Atena Roshan Fekr Katarzyna Radecka Zeljko Zilic |
author_facet | Majid Janidarmian Atena Roshan Fekr Katarzyna Radecka Zeljko Zilic |
author_sort | Majid Janidarmian |
collection | DOAJ |
description | Sensor-based motion recognition integrates the emerging area of wearable sensors with novel machine learning techniques to make sense of low-level sensor data and provide rich contextual information in a real-life application. Although Human Activity Recognition (HAR) problem has been drawing the attention of researchers, it is still a subject of much debate due to the diverse nature of human activities and their tracking methods. Finding the best predictive model in this problem while considering different sources of heterogeneities can be very difficult to analyze theoretically, which stresses the need of an experimental study. Therefore, in this paper, we first create the most complete dataset, focusing on accelerometer sensors, with various sources of heterogeneities. We then conduct an extensive analysis on feature representations and classification techniques (the most comprehensive comparison yet with 293 classifiers) for activity recognition. Principal component analysis is applied to reduce the feature vector dimension while keeping essential information. The average classification accuracy of eight sensor positions is reported to be 96.44% ± 1.62% with 10-fold evaluation, whereas accuracy of 79.92% ± 9.68% is reached in the subject-independent evaluation. This study presents significant evidence that we can build predictive models for HAR problem under more realistic conditions, and still achieve highly accurate results. |
first_indexed | 2024-04-11T22:52:46Z |
format | Article |
id | doaj.art-8ea2647e1074417780960b59178b6ec3 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-04-11T22:52:46Z |
publishDate | 2017-03-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-8ea2647e1074417780960b59178b6ec32022-12-22T03:58:32ZengMDPI AGSensors1424-82202017-03-0117352910.3390/s17030529s17030529A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity RecognitionMajid Janidarmian0Atena Roshan Fekr1Katarzyna Radecka2Zeljko Zilic3Electrical and Computer Engineering Department, McGill University, Montréal, QC H3A 0E9, CanadaElectrical and Computer Engineering Department, McGill University, Montréal, QC H3A 0E9, CanadaElectrical and Computer Engineering Department, McGill University, Montréal, QC H3A 0E9, CanadaElectrical and Computer Engineering Department, McGill University, Montréal, QC H3A 0E9, CanadaSensor-based motion recognition integrates the emerging area of wearable sensors with novel machine learning techniques to make sense of low-level sensor data and provide rich contextual information in a real-life application. Although Human Activity Recognition (HAR) problem has been drawing the attention of researchers, it is still a subject of much debate due to the diverse nature of human activities and their tracking methods. Finding the best predictive model in this problem while considering different sources of heterogeneities can be very difficult to analyze theoretically, which stresses the need of an experimental study. Therefore, in this paper, we first create the most complete dataset, focusing on accelerometer sensors, with various sources of heterogeneities. We then conduct an extensive analysis on feature representations and classification techniques (the most comprehensive comparison yet with 293 classifiers) for activity recognition. Principal component analysis is applied to reduce the feature vector dimension while keeping essential information. The average classification accuracy of eight sensor positions is reported to be 96.44% ± 1.62% with 10-fold evaluation, whereas accuracy of 79.92% ± 9.68% is reached in the subject-independent evaluation. This study presents significant evidence that we can build predictive models for HAR problem under more realistic conditions, and still achieve highly accurate results.http://www.mdpi.com/1424-8220/17/3/529human activity recognitionmachine learningsupervised classificationwearable sensorssensors heterogeneities |
spellingShingle | Majid Janidarmian Atena Roshan Fekr Katarzyna Radecka Zeljko Zilic A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition Sensors human activity recognition machine learning supervised classification wearable sensors sensors heterogeneities |
title | A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition |
title_full | A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition |
title_fullStr | A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition |
title_full_unstemmed | A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition |
title_short | A Comprehensive Analysis on Wearable Acceleration Sensors in Human Activity Recognition |
title_sort | comprehensive analysis on wearable acceleration sensors in human activity recognition |
topic | human activity recognition machine learning supervised classification wearable sensors sensors heterogeneities |
url | http://www.mdpi.com/1424-8220/17/3/529 |
work_keys_str_mv | AT majidjanidarmian acomprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT atenaroshanfekr acomprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT katarzynaradecka acomprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT zeljkozilic acomprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT majidjanidarmian comprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT atenaroshanfekr comprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT katarzynaradecka comprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition AT zeljkozilic comprehensiveanalysisonwearableaccelerationsensorsinhumanactivityrecognition |