Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks

Jeremiah Okai, Stylianos Paraschiakos, Marian Beekman, Arno Knobbe, Claudio Frederico Pinho Rebelo de Sá

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    Abstract

    Human Activity Recognition (HAR) is a growing field of research in biomedical engineering and it has many potential applications in the treatment and prevention of several diseases. Due to the recent advancement in technology, devices that collect position and orientation measurements (e.g. accelerometers and gyroscopes) are becoming ubiquitous. These measurements can then be used to train machine learning models for HAR. In this research, we propose one recurrent neural network architecture and a data augmentation approach for building robust and accurate models for HAR. We compared models with Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) layers. The proposed data augmentation approach was used to make the models robust to the cases where one or more sensors are missing. In this empirical study, we could also understand some relations between the ideal locations of the sensors in the participants and the types of activities performed. The proposed approaches were tested in the GOTOv dataset from a study which involved 35 participants performing 16 sedentary, ambulatory and lifestyle activities in a semi-structured environment. The results presented, clearly show that the models are able to detect these activities in a robust way.
    Original languageEnglish
    Title of host publication2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
    Place of PublicationPiscataway, NJ
    PublisherIEEE
    Number of pages6
    ISBN (Electronic)978-1-5386-1311-5
    ISBN (Print)978-1-5386-1312-2
    DOIs
    Publication statusPublished - 23 Jul 2019
    Event41st International Engineering in Medicine and Biology Conference, EMBC 2019: Biomedical Engineering Ranging from Wellness to Intensive Care Medicine - CityCube Berlin, Berlin, Germany
    Duration: 23 Jul 201927 Jul 2019
    Conference number: 41
    https://embc.embs.org/2019/

    Publication series

    NameAnnual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
    PublisherIEEE
    Number41
    Volume2019
    ISSN (Print)1557-170X
    ISSN (Electronic)1558-4615

    Conference

    Conference41st International Engineering in Medicine and Biology Conference, EMBC 2019
    Abbreviated titleEMBC
    CountryGermany
    CityBerlin
    Period23/07/1927/07/19
    Internet address

    Fingerprint

    Accelerometers
    Neural networks
    Biomedical engineering
    Recurrent neural networks
    Bioelectric potentials
    Gyroscopes
    Sensors
    Network architecture
    Learning systems
    Long short-term memory

    Cite this

    Okai, J., Paraschiakos, S., Beekman, M., Knobbe, A., & Pinho Rebelo de Sá, C. F. (2019). Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks. In 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); Vol. 2019, No. 41). Piscataway, NJ: IEEE. https://doi.org/10.1109/EMBC.2019.8857288
    Okai, Jeremiah ; Paraschiakos, Stylianos ; Beekman, Marian ; Knobbe, Arno ; Pinho Rebelo de Sá, Claudio Frederico. / Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks. 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Piscataway, NJ : IEEE, 2019. (Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); 41).
    @inproceedings{b40fe306d8df49c7a1cf6a7188eec9b8,
    title = "Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks",
    abstract = "Human Activity Recognition (HAR) is a growing field of research in biomedical engineering and it has many potential applications in the treatment and prevention of several diseases. Due to the recent advancement in technology, devices that collect position and orientation measurements (e.g. accelerometers and gyroscopes) are becoming ubiquitous. These measurements can then be used to train machine learning models for HAR. In this research, we propose one recurrent neural network architecture and a data augmentation approach for building robust and accurate models for HAR. We compared models with Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) layers. The proposed data augmentation approach was used to make the models robust to the cases where one or more sensors are missing. In this empirical study, we could also understand some relations between the ideal locations of the sensors in the participants and the types of activities performed. The proposed approaches were tested in the GOTOv dataset from a study which involved 35 participants performing 16 sedentary, ambulatory and lifestyle activities in a semi-structured environment. The results presented, clearly show that the models are able to detect these activities in a robust way.",
    author = "Jeremiah Okai and Stylianos Paraschiakos and Marian Beekman and Arno Knobbe and {Pinho Rebelo de S{\'a}}, {Claudio Frederico}",
    year = "2019",
    month = "7",
    day = "23",
    doi = "10.1109/EMBC.2019.8857288",
    language = "English",
    isbn = "978-1-5386-1312-2",
    series = "Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)",
    publisher = "IEEE",
    number = "41",
    booktitle = "2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)",
    address = "United States",

    }

    Okai, J, Paraschiakos, S, Beekman, M, Knobbe, A & Pinho Rebelo de Sá, CF 2019, Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks. in 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), no. 41, vol. 2019, IEEE, Piscataway, NJ, 41st International Engineering in Medicine and Biology Conference, EMBC 2019, Berlin, Germany, 23/07/19. https://doi.org/10.1109/EMBC.2019.8857288

    Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks. / Okai, Jeremiah ; Paraschiakos, Stylianos; Beekman, Marian; Knobbe, Arno; Pinho Rebelo de Sá, Claudio Frederico.

    2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Piscataway, NJ : IEEE, 2019. (Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); Vol. 2019, No. 41).

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    TY - GEN

    T1 - Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks

    AU - Okai, Jeremiah

    AU - Paraschiakos, Stylianos

    AU - Beekman, Marian

    AU - Knobbe, Arno

    AU - Pinho Rebelo de Sá, Claudio Frederico

    PY - 2019/7/23

    Y1 - 2019/7/23

    N2 - Human Activity Recognition (HAR) is a growing field of research in biomedical engineering and it has many potential applications in the treatment and prevention of several diseases. Due to the recent advancement in technology, devices that collect position and orientation measurements (e.g. accelerometers and gyroscopes) are becoming ubiquitous. These measurements can then be used to train machine learning models for HAR. In this research, we propose one recurrent neural network architecture and a data augmentation approach for building robust and accurate models for HAR. We compared models with Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) layers. The proposed data augmentation approach was used to make the models robust to the cases where one or more sensors are missing. In this empirical study, we could also understand some relations between the ideal locations of the sensors in the participants and the types of activities performed. The proposed approaches were tested in the GOTOv dataset from a study which involved 35 participants performing 16 sedentary, ambulatory and lifestyle activities in a semi-structured environment. The results presented, clearly show that the models are able to detect these activities in a robust way.

    AB - Human Activity Recognition (HAR) is a growing field of research in biomedical engineering and it has many potential applications in the treatment and prevention of several diseases. Due to the recent advancement in technology, devices that collect position and orientation measurements (e.g. accelerometers and gyroscopes) are becoming ubiquitous. These measurements can then be used to train machine learning models for HAR. In this research, we propose one recurrent neural network architecture and a data augmentation approach for building robust and accurate models for HAR. We compared models with Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) layers. The proposed data augmentation approach was used to make the models robust to the cases where one or more sensors are missing. In this empirical study, we could also understand some relations between the ideal locations of the sensors in the participants and the types of activities performed. The proposed approaches were tested in the GOTOv dataset from a study which involved 35 participants performing 16 sedentary, ambulatory and lifestyle activities in a semi-structured environment. The results presented, clearly show that the models are able to detect these activities in a robust way.

    U2 - 10.1109/EMBC.2019.8857288

    DO - 10.1109/EMBC.2019.8857288

    M3 - Conference contribution

    SN - 978-1-5386-1312-2

    T3 - Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

    BT - 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

    PB - IEEE

    CY - Piscataway, NJ

    ER -

    Okai J, Paraschiakos S, Beekman M, Knobbe A, Pinho Rebelo de Sá CF. Building robust models for Human Activity Recognition from raw accelerometers data using Gated Recurrent Units and Long Short Term Memory Neural Networks. In 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Piscataway, NJ: IEEE. 2019. (Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); 41). https://doi.org/10.1109/EMBC.2019.8857288