Comparing the Accuracy of Multiple Commercial Wearable Devices: A Method

To maintain a good fitness level, performing a minimum daily physical activity is important. Many people think that they have already met the suggested guidelines. However, most of them were wrong when they quantify their activity level using a personal activity tracker device. This problem indicates the need for personal physical activity tracker or wearable device to maintain the fitness level. However, as wide-range wearable devices available today, it is difficult to find the most suitable one regarding the price and quality. This paper proposes a study to explore the association between the quality and the price of the devices. The quality is evaluated by comparing the accuracy of the heart rate sensor and step counter of several wearable devices with the ground truth measurements. The test result is then analyzed using a series of statistical methods including, a paired t-test, Pearson’s correlation test, and Bland-Altman plot. Both t-test and correlation coefficient is used to plot the association between quality and price.