Wearable Sensors and Smartphone Apps as Pedometers in eHealth: a Comparative Accuracy, Reliability and User Evaluation

Abstract

Research in sensor technology has provided the scientific community with advanced sensors with higher speed and lower costs, promoting the manufacturing of more efficient, comfortable and affordable wearable devices coupled with pervasive smartphone app usage. A prominent use of wearables and apps is to leverage built-in accelerometers to estimate the user’s steps as a measure of physical activity in an eHealth and well-being context. This study aims to evaluate the accuracy, reliability and user preferences of nine prominent devices and apps in a trial with 33 healthy adults, in natural walking conditions. Based on sales, functions and capabilities we selected 5 wearable devices (Jawbone UP3, Jawbone UP24, Fitbit Charge HR, Fitbit Zip and Microsoft Band) and 4 smartphone apps (Google Fit, Accupedo, Noom Walk and Runtastic). The experiment consisted of three walking tasks, from short to long walking, performed in a wide hallway instead of a treadmill. Steps were counted using a tally counter and measurements were gathered after each task, in software developed for analysis. Results also include data loss and reliability, user preferences and evaluation of trackers through a tailored questionnaire. This generation of devices showed increased accuracy, with error rates of around 5%, mostly in long trials, and small loss of data. Applications are generally showing less accuracy. The study shows that there are trackers in the market that combine high accuracy, reliability, features desired by users and affordable prices, which make them suitable for eHealth applications.