Reliability and validity of estimated angles information assessed using inertial measurement unit-based motion sensors

Abstract

BACKGROUND: Inertial measurement unit (IMU)-based motion sensors are affordable, and their use is appropriate for rehabilitation. However, regarding the accuracy of estimated angle information obtained from this sensor, it is reported that it is likely affected by velocity. OBJECTIVE: The present study investigated the reliability and validity of the angle information obtained using IMU-based sensors compared with a three-dimensional (3D) motion analyzer. METHODS: The Euler angle obtained using the 3D motion analyzer and the angle obtained using the IMU-based sensor (IMU angle) were compared. Reliability was assessed by comparing the Bland–Altman analysis, intra-class correlation coefficient (ICC) (1,1), and cross-correlation function. The root mean square (RMS) error, ICC (2,1), and cross-correlation function were used to compare data on the Euler and IMU angles to evaluate the validity. RESULTS: Regarding reliability, the Bland–Atman analysis indicated no fixed or proportional bias in the angle measurements. The measurement errors ranged from 0.2° to 3.2°. In the validity, the RMS error ranged from 0.3° to 2.2°. The ICCs (2,1) were 0.9. The cross-correlation functions were >0.9, which indicated a high degree of agreement. CONCLUSION: The IMU-based sensor had a high reliability and validity. The IMU angle may be used in rehabilitation.