010 20 30 40 50 60
AAMI ISO 81060-2
Mean error: -2.1 mmHg
Stdev error: 6.8 mmHg
Fig. 1: Evolution of blood pressure in a healthy volunteer during two
months. Grey boxes depict diastolic oscillometric measurements (±8
mmHg), green dots are estimates from Aktiia OBPM algorithms. Only
the first measurement at Day 0 was used to initialize estimates.
Measurement up to two weeks after initialization
Measurement up to two months after initialization
Fig. 2: Study Bland-Altman plot, with measurements performed until
two weeks (green boxes), and two months after initialization (red dots).
Abstract— The performance of the Aktiia OBPM algorithms
to measure blood pressure at the wrist was investigated in this
study. For two months, six volunteers recorded blood pressure
values at the arm using a brachial cuff. Simultaneously, optical
signals at the wrist were recorded using off-the-shelf PPG
sensors. At the end of the study, the optical signals were
processed by the Aktiia OBPM algorithms to generate blood
pressure estimations. The algorithms were initialized using the
first brachial blood pressure value recorded at the inclusion day.
The blood pressure readings estimated by the Aktiia OBPM
algorithms fell within the AAMI/ISO81060-2 requirements up to
two months after the initialization procedure.
Hypertension is the largest epidemic ever known to mankind,
affecting 30% of the adult population, and killing 9 million
people per year around the world. In order to reduce the
worldwide impact of hypertension, the early detection and the
optimized management of elevated blood pressure are to be
rethought. Based on 15 years of research  Aktiia SA is
preparing the deployment of a bracelet that will allow the
measurement of blood pressure around the clock, and that will
easily integrate into people’s life by reducing the use of
inflation cuffs . Aktiia SA is currently optimizing its library
of OBPM algorithms that will process PPG signals acquired
by off-the-shelf optical sensors at the wrist. The current study
investigated the performance of Aktiia OBPM algorithms,
demonstrating their stability after initialization.
For two months, six healthy volunteers recorded simultaneous
blood pressure readings at the brachial artery (OMRON M6),
and contralateral reflective PPG signals at the wrist via an
off-the-shelf optical sensor (OSRAM SFH 7072). Isometric
leg extensions were also performed in order to induce large
blood pressure variations. The PPG signals recorded during
the study were retrospectively analyzed by the Aktiia OBPM
algorithms. For each volunteer, initialization of the algorithms
was performed based on the first blood pressure reading at the
inclusion day (used to re-offset Aktiia OBPM estimations).
Fig. 1 illustrates an example of two-month evolution of
oscillometric diastolic blood pressure measurements (study
reference) compared to cuffless estimates from Aktiia OBPM
algorithms. Table 1 summarizes overall performances in
terms of AAMI/ISO81060-2 requirements , and Fig. 2
illustrates Bland-Altman plot for the entire study.
TABLE I. OVERALL PERFORMANCE OF AKTIIA OBPM ALGORITHMS
Up to two weeks after
Up to two months after
IV. DISCUSSION & CONCLUSION
For this study, Aktiia OBPM algorithms were able to calculate
blood pressure estimates falling within the requirements of
AAMI/ISO81060-2 up to two months after their initialization.
 J. Solà et al, IEEE Pulse 2018, doi: 10.1109/MPUL.2018.2856960.
 Aktiia SA website 2019, www.aktiia.com
 AAMI/ISO, 2018, AAMI/ISO 81060-2:2018.
Aktiia Bracelet: Monitoring of Blood Pressure
using Off-the-shelf Optical Sensors
J. Solà, A. Vybornova, S. Fallet, O. Grossenbacher, B. De Marco, E. Olivero, N. Siutryk, V. Chapuis and M. Bertschi
CONFIDENTIAL. Limited circulation. For review only.
Manuscript 265 submitted to 2019 41st Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC). Received January 23, 2019.