Sensitivity analysis of an inertial calibration method: Influence of leg position on 3D knee kinematics

Background

Inertial systems are increasingly used to analyze human motion, especially for gait analyses and in clinical settings. Calibration methods for these systems are designed for ease of implementation, and previous studies have shown that they can provide accurate knee kinematics in the sagittal plane. However, the reason of their lack of accuracy in the other planes (i.e., transverse and frontal) remain unknown.

Research question

This study aimed to evaluate the sensitivity of one posture of a double-pose calibration method to analyse 3D knee kinematics during gait with two magnetic inertial measurement units (MIMU). This method consists of a standing posture and a posture with the leg stretched forward in the sagittal plane, which together define the sagittal plane. Our hypothesis was that a change in the definition of the sagittal plane during the calibration process was likely to affect the assessment of knee kinematics in the frontal and transverse planes.

Methods

Ten healthy participants wearing the KneeKG system and two MIMU completed the calibration process in five different leg positions (0°, 3°, 5°, 10° or 15° from the sagittal plane) for the second calibration posture. After static calibration, the participants walked on an instrumented treadmill at a speed of 1.1 m/s and 3D knee kinematics were calculated using the five different calibration conditions.

Results

Mean absolute difference (MAD) between the swing-phase peak value of the curve corresponding to the leg in the sagittal plane (0° shift from this plane) when performing the second calibration posture and each of the other curves was 0.20–0.46° for knee flexion, 1.67–2.90° for adduction, and 0.72–1.46° for external rotation. MAD of the swing-phase peak value in the frontal plane was correlated (R²=0.81) with the angulation of the femur in the sagittal plane during calibration.

Significance

An angular shift from the sagittal plane when performing a double-pose calibration method induces a minimal influence on the knee flexion/extension but larger influences on secondary knee motions.