Self-Compensating Angle Estimation Algorithms for Quantitative Evaluation of Gait Kinematics in Physiotherapy

Abstract

This thesis presents the “easyWearable Gait Monitor,” an IMU-based system with a low-cost microcontroller for physiotherapy tracking. It incorporates proposed algorithms, including a novel gyroscope-based flexion/extension angle estimation with 99.2% CORR, and an accelerometer-based vertical offset correction method with 96.4% CORR. A yaw-tilt correction method with 97.8% accuracy to overcome sensor misplacement for global abduction/adduction angle estimation is also proposed. All algorithms are validated against a low-cost marker-based MoCap system with low RMSE of 0.83°.