Body-Worn IMU-Based Human Hip and Knee Kinematics Estimation during Treadmill Walking

Area of Science:

• Biomechanics
• Wearable Technology
• Human Motion Analysis

Background:

• Inertial measurement unit (IMU)-based human joint angle estimation often assumes full joint mobility.
• Human walking presents limited joint motion, causing unobservability issues for magnetometer-free IMU methods.
• Existing techniques struggle with accuracy during gait due to specific kinematic constraints.

Purpose of the Study:

• To explore unobservability conditions during human walking for IMU-based motion analysis.
• To expand a previous IMU method for estimating hip angles alongside knee angles.
• To develop a magnetometer-free IMU approach for skeletal pose estimation during gait.

Main Methods:

• Developed an IMU-based method to estimate human hip and knee angles during walking.
• Evaluated the method in a 12-subject study, comparing results against an optical motion capture system.
• Investigated error sources, including IMU heading drift and soft tissue perturbations.

Main Results:

• Achieved comparable accuracy to state-of-the-art methods for knee flexion/extension (7.87° RMSE) and hip flexion/extension (3.70° RMSE).
• Observed higher errors in hip internal/external rotation (6.27° RMSE) due to heading drift and complex hip kinematics.
• Identified soft tissue effects causing overestimation during stance and underestimation during swing for knee angles.

Conclusions:

• The proposed method offers a novel approach to magnetic-blind, kinematic-only IMU-based skeletal pose estimation.
• Addresses observability challenges in human walking, a task with degenerate kinematics.
• Provides insights into error sources and potential corrections for IMU gait analysis.