Inertial motion capture validation of 3D knee kinematics at various gait speed on the treadmill with a double-pose calibration

•Inertial motion capture knee angles compared favorably to optical motion capture.•Inertial motion capture calibrated with a double-pose showed an acceptable error.•3D knee kinematics can be obtained independently from inertial motion capture.•Increases in gait speed on the treadmill show progressive changes in knee angles.•Knee kinematics evolution with gait speed was detected similarly by both systems. Inertial motion capture (IMC) is rapidly gaining in popularity to evaluate gait in clinical settings. Previous examinations of IMC knee kinematics were often limited to the sagittal plane and IMC calibration has not been thoroughly investigated. The objective was to validate IMC 3D knee kinematics calibrated with a double-pose during gait with reference to optical motion capture (OMC). The hypotheses are that IMC can estimate adequately knee kinematics and that both systems will detect similarly the changes with gait speed. Twenty-four healthy participants walked on the treadmill at gait speed of 0.6, 0.8, 1.0 and 1.2 m/s. Knee kinematics were obtained simultaneously with two magnetic and inertial measurement units and passive markers fixed on the KneeKG system. OMC was calibrated with a functional anatomical approach and the IMC with a double-pose. Root mean square differences of the two systems yielded 3-6° for knee flexion, adduction and external rotation. Knee kinematics were more similar during the stance phase than the swing phase. Gait speed showed a significant progressive effect on the three knee angles that was similarly detected by the two systems. IMC 3D knee kinematics can be obtained independently with a simple calibration and only two magnetic and inertial measurement units at an acceptable level of error especially during stance.