Inertial measurement units-based two-dimensional joint angle estimation method of lower limb for running using a camera tracking calibration

Abstract

The analysis of joint angles is used to assess gait and running performance. Inertial measurement units (IMUs) are compact, cost-effective devices that can be attached to the body to perform simple motion analysis. However, the use of IMUs necessitates the resolution of two key issues: compensating for drift and mounting errors. The proposed method employs an extended Kalman filter (EKF) to compensate for drift error and position data from a camera to compensate for mounting error. Joint angles are calculated using the data compensated by the EKF. The proposed method calculates two-dimensional joint angles and is suitable for the performance analysis of gait and running. A measurement experiment was conducted with a runner to validate the proposed method. The IMUs were attached at arbitrary positions on body segments and misalignment with body axis, and position markers were attached for the camera-based joint angle calculation. The experimental protocol comprised four phases: static, acceleration (including calibration period), measurement, and deceleration. The compensation of IMUs mounting error was performed using actual motion data during the calibration period measured by the camera. The running velocity in the measurement phase was 25 km/h. The experimental results demonstrated that the drift and mounting errors due to the IMU measurements were corrected. The accuracy of joint angle measurements was demonstrated by comparing the joint angle results obtained with the method using only a camera, showing the effectiveness of proposed method. Although the proposed method uses both IMUs and a camera, the data from the camera are only required for the calibration period, which is a short part of the experiment. Therefore, the method can be used to improve the accuracy of simple IMU measurements and is expected to be expanded to outdoor motion measurements.