Methods for validating the performance of wearable motion-sensing devices under controlled conditions

This paper presents validation methods for assessing the accuracy and precision of motion-sensing device (i.e. accelerometer) measurements. The main goals of this paper were to assess the accuracy and precision of these measurements against a gold standard, to determine if differences in manufacturing and assembly significantly affected device performance and to determine if measurement differences due to manufacturing and assembly could be corrected by applying certain post-processing techniques to the measurement data during analysis. In this paper, the validation of a posture and activity detector (PAD), a device containing a tri-axial accelerometer, is described. Validation of the PAD devices required the design of two test fixtures: a test fixture to position the device in a known orientation, and a test fixture to rotate the device at known velocities and accelerations. Device measurements were compared to these known orientations and accelerations. Several post-processing techniques were utilized in an attempt to reduce variability in the measurement error among the devices. In conclusion, some of the measurement errors due to the inevitable differences in manufacturing and assembly were significantly improved (p < 0.01) by these post-processing techniques.