Accuracy of the Motorola Defy smartphone for assessing physical activity and distance travelled

Purpose: The onboard accelerometers and global positioning system (GPS) receivers on modern smartphones offer considerable promise for the unobtrusive tracking of physical activity (PA) in population research; however, the accuracy of smartphone sensors for assessing PA is uncertain. The purpose of this study was to compare estimates of PA and distance traveled between a popular smartphone and standalone accelerometer and GPS devices.

Methods: A total of 13 men and 19 women aged 19-47 years were fitted with a Motorola Defy smartphone (in a vertical holster), an Actigraph GT3X+ accelerometer, and a Qstarz BT-Q1000XT GPS receiver on an elastic belt around the waist. Accelerometer and GPS data were accessed from the Defy sensors using a custom application (ActiPal). Participants moved around a 400 m track at four self-selected speeds (slow walk, fast walk, slow run, fast run) before completing six lengths of lateral sidesteps between two points 20 m apart.

Results: Compared with the GT3X+, the Defy overestimated the vector magnitude of the raw acceleration forces by 11.9 ± 8.8%, 16.6 ± 12.1%, and 17.6 ± 11.2% for slow jogging, fast jogging, and side stepping (respectively). Compared with the BT-Q1000XT, the Defy overestimated total distance traveled by 26.9 ± 19.0%, 27.3 ± 19.5%, and 18.6 ± 14.9% for slow walking, fast walking, and slow jogging.

Conclusions: The significant differences between devices indicate that care should be taken when using onboard smartphone sensors to collect PA and positional data for research purposes.