Toward Automatic Cardiovascular and Respiratory Assessment Using Automatic 6-Minute Walking Test

The aim of this study was to investigate suitable cardiorespiratory features for automatic six minute walking test for home monitoring using inertial measurement units and electrocardiography by recognizing walking and standing, estimating the intensity of the walking and assessing the physi-ological response to the walking. The walking was recognized from gyroscope and accelerometer signals from a wrist with accuracy of 84% using XGBoost. To estimate the functional capacity, heart rate, heart rate recovery time and respiratory rate were calculated from electrocardiogram and accelerometer signal, respectively. To evaluate the physiological response to the exercise, several features to describe the walking intensity were calculated from the accelerometer signal from the chest. Feature that best correlated with the heart rate, heart rate recovery time and respiratory rate was walking frequency with R value of 0.65, 0.47 and 0.55, respectively and relative change of 31 %. This was a pilot study to show that motion sensors and electrocardiography could be used to perform an automatic six minute walking test which could be used for example in assessing functional capacity of heart failure patients.