Deep activity recognition models with triaxial accelerometers

Mohammad Abu Alsheikh, Ahmed Selim, Dusit Niyato, Linda Doyle, Shaowei Lin, Hwee Pink Tan

Research output: A Conference proceeding or a Chapter in BookConference contribution

39 Citations (Scopus)

Abstract

Despite the widespread installation of accelerometers in almost all mobile phones and wearable devices, activity recognition using accelerometers is still immature due to the poor recognition accuracy of existing recognition methods and the scarcity of labeled training data. We consider the problem of human activity recognition using triaxial accelerometers and deep learning paradigms. This paper shows that deep activity recognition models (a) provide better recognition accuracy of human activities, (b) avoid the expensive design of handcrafted features in existing systems, and (c) utilize the massive unlabeled acceleration samples for unsupervised feature extraction. We show substantial recognition improvement on real world datasets over state-of-the-art methods of human activity recognition using triaxial accelerometers.

Original languageEnglish
Title of host publicationAAAI Conference on Artificial Intelligence
Subtitle of host publicationWorkshop on Artificial Intelligence Applied to Assistive Tecnologies and Smart Environments
Place of PublicationPheonix, United States
PublisherAI Access Foundation
Pages8-13
Number of pages6
VolumeWS-16-01 - WS-16-15
ISBN (Electronic)9781577357599
ISBN (Print)9781577357599
Publication statusPublished - 2016
Externally publishedYes
Event30th AAAI Conference on Artificial Intelligence, AAAI 2016 - Phoenix, United States
Duration: 12 Feb 201613 Feb 2016

Publication series

NameAAAI Workshop - Technical Report
VolumeWS-16-01 - WS-16-15

Conference

Conference30th AAAI Conference on Artificial Intelligence, AAAI 2016
CountryUnited States
CityPhoenix
Period12/02/1613/02/16

Fingerprint

Accelerometers
Mobile phones
Feature extraction

Cite this

Abu Alsheikh, M., Selim, A., Niyato, D., Doyle, L., Lin, S., & Tan, H. P. (2016). Deep activity recognition models with triaxial accelerometers. In AAAI Conference on Artificial Intelligence: Workshop on Artificial Intelligence Applied to Assistive Tecnologies and Smart Environments (Vol. WS-16-01 - WS-16-15, pp. 8-13). [WS-16-01] (AAAI Workshop - Technical Report; Vol. WS-16-01 - WS-16-15). Pheonix, United States: AI Access Foundation.
Abu Alsheikh, Mohammad ; Selim, Ahmed ; Niyato, Dusit ; Doyle, Linda ; Lin, Shaowei ; Tan, Hwee Pink. / Deep activity recognition models with triaxial accelerometers. AAAI Conference on Artificial Intelligence: Workshop on Artificial Intelligence Applied to Assistive Tecnologies and Smart Environments. Vol. WS-16-01 - WS-16-15 Pheonix, United States : AI Access Foundation, 2016. pp. 8-13 (AAAI Workshop - Technical Report).
@inproceedings{2e3f5a178ea442d89773b1c3698cb1a6,
title = "Deep activity recognition models with triaxial accelerometers",
abstract = "Despite the widespread installation of accelerometers in almost all mobile phones and wearable devices, activity recognition using accelerometers is still immature due to the poor recognition accuracy of existing recognition methods and the scarcity of labeled training data. We consider the problem of human activity recognition using triaxial accelerometers and deep learning paradigms. This paper shows that deep activity recognition models (a) provide better recognition accuracy of human activities, (b) avoid the expensive design of handcrafted features in existing systems, and (c) utilize the massive unlabeled acceleration samples for unsupervised feature extraction. We show substantial recognition improvement on real world datasets over state-of-the-art methods of human activity recognition using triaxial accelerometers.",
keywords = "Activity recognition, deep learning, feature learning, accelerometers",
author = "{Abu Alsheikh}, Mohammad and Ahmed Selim and Dusit Niyato and Linda Doyle and Shaowei Lin and Tan, {Hwee Pink}",
year = "2016",
language = "English",
isbn = "9781577357599",
volume = "WS-16-01 - WS-16-15",
series = "AAAI Workshop - Technical Report",
publisher = "AI Access Foundation",
pages = "8--13",
booktitle = "AAAI Conference on Artificial Intelligence",
address = "United States",

}

Abu Alsheikh, M, Selim, A, Niyato, D, Doyle, L, Lin, S & Tan, HP 2016, Deep activity recognition models with triaxial accelerometers. in AAAI Conference on Artificial Intelligence: Workshop on Artificial Intelligence Applied to Assistive Tecnologies and Smart Environments. vol. WS-16-01 - WS-16-15, WS-16-01, AAAI Workshop - Technical Report, vol. WS-16-01 - WS-16-15, AI Access Foundation, Pheonix, United States, pp. 8-13, 30th AAAI Conference on Artificial Intelligence, AAAI 2016, Phoenix, United States, 12/02/16.

Deep activity recognition models with triaxial accelerometers. / Abu Alsheikh, Mohammad; Selim, Ahmed; Niyato, Dusit; Doyle, Linda; Lin, Shaowei; Tan, Hwee Pink.

AAAI Conference on Artificial Intelligence: Workshop on Artificial Intelligence Applied to Assistive Tecnologies and Smart Environments. Vol. WS-16-01 - WS-16-15 Pheonix, United States : AI Access Foundation, 2016. p. 8-13 WS-16-01 (AAAI Workshop - Technical Report; Vol. WS-16-01 - WS-16-15).

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Deep activity recognition models with triaxial accelerometers

AU - Abu Alsheikh, Mohammad

AU - Selim, Ahmed

AU - Niyato, Dusit

AU - Doyle, Linda

AU - Lin, Shaowei

AU - Tan, Hwee Pink

PY - 2016

Y1 - 2016

N2 - Despite the widespread installation of accelerometers in almost all mobile phones and wearable devices, activity recognition using accelerometers is still immature due to the poor recognition accuracy of existing recognition methods and the scarcity of labeled training data. We consider the problem of human activity recognition using triaxial accelerometers and deep learning paradigms. This paper shows that deep activity recognition models (a) provide better recognition accuracy of human activities, (b) avoid the expensive design of handcrafted features in existing systems, and (c) utilize the massive unlabeled acceleration samples for unsupervised feature extraction. We show substantial recognition improvement on real world datasets over state-of-the-art methods of human activity recognition using triaxial accelerometers.

AB - Despite the widespread installation of accelerometers in almost all mobile phones and wearable devices, activity recognition using accelerometers is still immature due to the poor recognition accuracy of existing recognition methods and the scarcity of labeled training data. We consider the problem of human activity recognition using triaxial accelerometers and deep learning paradigms. This paper shows that deep activity recognition models (a) provide better recognition accuracy of human activities, (b) avoid the expensive design of handcrafted features in existing systems, and (c) utilize the massive unlabeled acceleration samples for unsupervised feature extraction. We show substantial recognition improvement on real world datasets over state-of-the-art methods of human activity recognition using triaxial accelerometers.

KW - Activity recognition

KW - deep learning

KW - feature learning

KW - accelerometers

UR - http://www.scopus.com/inward/record.url?scp=85017617585&partnerID=8YFLogxK

UR - http://www.mendeley.com/research/planning-scheduling-monitoring-airport-surface-operations

M3 - Conference contribution

SN - 9781577357599

VL - WS-16-01 - WS-16-15

T3 - AAAI Workshop - Technical Report

SP - 8

EP - 13

BT - AAAI Conference on Artificial Intelligence

PB - AI Access Foundation

CY - Pheonix, United States

ER -

Abu Alsheikh M, Selim A, Niyato D, Doyle L, Lin S, Tan HP. Deep activity recognition models with triaxial accelerometers. In AAAI Conference on Artificial Intelligence: Workshop on Artificial Intelligence Applied to Assistive Tecnologies and Smart Environments. Vol. WS-16-01 - WS-16-15. Pheonix, United States: AI Access Foundation. 2016. p. 8-13. WS-16-01. (AAAI Workshop - Technical Report).