Motion2Vector: unsupervised learning in human activity recognition using wrist-sensing data

Lu Bai, Chris Yeung, Christos Efstratiou, Moyra Chikomo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Downloads (Pure)

Abstract

With the increasing popularity of consumer wearable devices augmented with sensing capabilities (smart bands, smart watches), there is a significant focus on extracting meaningful information about human behaviour through large scale real-world wearable sensor data. The focus of this work is to develop techniques to detect human activities, utilis-ing a large dataset of wearable data where no ground truth has been produced on the actual activities performed. We propose a deep learning variational auto encoder activity recognition model - Motion2Vector. The model is trained using large amounts of unlabelled human activity data to learn a representation of a time period of activity data. The learned activity representations can be mapped into an embedded activity space and grouped with regards to the nature of the activity type. In order to evaluate the proposed model, we have applied our method on public dataset - The Heterogeneity Human Activity Recognition (HHAR) dataset. The results showed that our method can achieve improved result over the HHAR dataset. In addition, we have collected our own lab-based activity dataset. Our experimental results show that our system achieves good accuracy in detecting such activities, and has the potential to provide additional insights in understanding the real-world activity in the situations where there is no ground truth available.

Original languageEnglish
Title of host publicationUbiComp/ISWC 2019 - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers
Place of PublicationNew York
PublisherAssociation for Computing Machinery, Inc
Pages537-542
Number of pages6
ISBN (Print)978-1-4503-6869-8
DOIs
Publication statusPublished - 9 Sep 2019
Eventthe 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2019 ACM International Symposium - QEII Centre, London, United Kingdom
Duration: 9 Sep 201913 Sep 2019
http://ubicomp.org/ubicomp2019/

Conference

Conferencethe 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2019 ACM International Symposium
CountryUnited Kingdom
CityLondon
Period9/09/1913/09/19
Internet address

Keywords

  • Deep learning
  • Human activity recognition
  • Variational auto encoder

Fingerprint Dive into the research topics of 'Motion2Vector: unsupervised learning in human activity recognition using wrist-sensing data'. Together they form a unique fingerprint.

  • Cite this

    Bai, L., Yeung, C., Efstratiou, C., & Chikomo, M. (2019). Motion2Vector: unsupervised learning in human activity recognition using wrist-sensing data. In UbiComp/ISWC 2019 - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers (pp. 537-542). Association for Computing Machinery, Inc. https://doi.org/10.1145/3341162.3349335