Using Inertial Measurement Unit Sensor Single Axis Rotation Angles for Knee and Hip Flexion Angle Calculations During Gait

Abstract

Background: Hip and knee flexion joint motions are frequently examined in clinical practice using camera based motion capture (CBMC) systems; however, these systems require elaborate setups and dedicated space. Inertial measurement unit (IMU) based systems avoid these disadvantages but require validation before widespread adoption. Moreover, it is important for clinical practice to determine the stability of these systems for prolonged evaluation periods. The purpose of this study was to assess the validity of a three-sensor inertial measurement unit system for calculating hip and knee flexion angles during gait by comparing with a gold standard CBMC system. Validity was also examined before and after a treadmill walking session. Methods: Twenty healthy participants were tested. Twenty seconds of gait at preferred walking speed were analyzed before and after thirty-two minutes of treadmill walking using previously validated CBMC methods and with a custom IMU model. Measurement validity for the IMU system was evaluated using Bland & Altman 95 percent limits of agreement, linear regression, mean absolute error and root mean square error. The effects of a measurement zeroing calibration strategy were also investigated. Results: Strong measurement agreement was observed for both hip and knee flexion angles, although overall agreement for the hip exceeded that for the knee. Linear regressions between the datasets for each participant illustrated strong (> 0.94) relationships between IMU and CBMC measurements. More significant changes between timepoints were observed for the knee than for the hip. Error values were generally reduced when zeroing calibration was implemented. Conclusion:The IMU system presented in this study is a convenient and accessible technique to measure joint angles. The protocol described in the current study can be easily applied in the clinical setting for evaluation of clinical populations. Additional development work on sensor placement and calibration methods may further increase the accuracy of such methods.