IoT Technologies for HealthCare. 6th EAI International Conference, HealthyIoT 2019, Braga, Portugal, December 4–6, 2019, Proceedings

Research Article

Towards a Smartwatch for Cuff-Less Blood Pressure Measurement Using PPG Signal and Physiological Features

Download234 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-42029-1_5,
        author={Franck Mouney and Teodor Tiplica and Magid Hallab and Mickeal Dinomais and Jean-Baptiste Fasquel},
        title={Towards a Smartwatch for Cuff-Less Blood Pressure Measurement Using PPG Signal and Physiological Features},
        proceedings={IoT Technologies for HealthCare. 6th EAI International Conference, HealthyIoT 2019, Braga, Portugal, December 4--6, 2019, Proceedings},
        proceedings_a={HEALTHYIOT},
        year={2020},
        month={6},
        keywords={Photoplethysmogram (PPG) Blood pressure (BP) FFT Lasso IoT},
        doi={10.1007/978-3-030-42029-1_5}
    }
    
  • Franck Mouney
    Teodor Tiplica
    Magid Hallab
    Mickeal Dinomais
    Jean-Baptiste Fasquel
    Year: 2020
    Towards a Smartwatch for Cuff-Less Blood Pressure Measurement Using PPG Signal and Physiological Features
    HEALTHYIOT
    Springer
    DOI: 10.1007/978-3-030-42029-1_5
Franck Mouney,*, Teodor Tiplica1, Magid Hallab2, Mickeal Dinomais1, Jean-Baptiste Fasquel1
  • 1: Angers University
  • 2: Axelife SAS
*Contact email: mouney.franck@gmail.com

Abstract

The context of this work concerns the development of a connected smartwatch for the continuous daily monitoring of physiological parameters to prevent cardiovascular diseases, and for the follow-up of the efficiency of treatments, against hypertension for example. This paper focuses on a particular parameter, the blood pressure (BP), to be automatically measured from the Photoplethysmogram (PPG) signal, to be acquired using a smartwatch. The proposed method is based on the automatic pulse wave detection from the PPG signal. Then, using the Lasso algorithm, a relation has been established between the blood pressure and the spectral representation of the normalized pulse wave, combined with other physiological information (age, body mass index and hear rate). The proposed method has been evaluated on a recent large public database of 219 subjects, covering a large range of ages (20–89), body mass indices and of blood pressures. Experimental results show acceptable performances in terms of accuracy. Compared to a recent related work depicting a slightly lower estimation error, a strength of our approach regards its robustness with respect to the signal quality, this being crucial for a use in daily routine in real IoT conditions, as it is the case in this context of smartwatch.