Cardiopulmonary Resuscitation Quality Parameters from Inertial Sensor Data using Differential Evolution Fitting of Sinusoids

In this paper, we present a robust sinusoidal model fitting method based on the Differential Evolution (DE) algorithm for determining cardiopulmonary resuscitation (CPR) quality-parameters - naming chest compression frequency and depth - as measured by an inertial sensor placed at the wrist. Once included into a smartphone or smartwatch app, our proposed algorithm will enable laypersons to improve cardiopulmonary resuscitation (as part of a continuous closed-loop support-system). By evaluating the sensitivity of the model with data recorded by a Laerdal Resusci Anne mannequin as reference standard, a low variance for compression frequency of +-2.7 cpm (2.5 %) has been found for the sensor placed at the wrist, making this previously not evaluated position a suitable alternative to the typical smartphone placement in the hand.