Knee Angle Estimation with Dynamic Calibration Using Inertial Measurement Units for Running.

Area of Science:

• Biomechanics
• Human Movement Analysis
• Wearable Sensor Technology

Background:

• Accurate knee flexion angle measurement is crucial for gait analysis, especially during running, a high-risk activity for knee injuries.
• Laboratory-based optical motion-capture systems provide accuracy but limit real-world running gait studies.
• Existing wearable sensor methods often require complex setups like sensor-to-segment assumptions or specific calibration poses.

Observation:

• This study investigated the use of shank and thigh inertial sensors (accelerometer and gyroscope) to estimate knee flexion angle during running.
• Three different filtering algorithms were evaluated without relying on magnetometers, sensor-to-segment assumptions, or specific calibration poses.
• Data from a single participant across four treadmill speeds were compared against a Vicon optical motion-capture system.

Findings:

• The developed filtering algorithms, particularly a complementary filter approach, accurately estimated knee flexion angle during running.
• Root-mean-square errors were approximately three degrees, well within the acceptable five-degree limit for clinical gait analysis.
• The proposed method successfully estimates knee flexion using only accelerometer and gyroscope data, eliminating the need for complex sensor mounting or calibration.

Implications:

• This research provides a viable, sensor-based method for estimating knee flexion angle in real-world running scenarios.
• The findings support the use of inertial sensors for accessible and accurate gait analysis, potentially aiding in injury prevention and rehabilitation.
• The complementary filter approach demonstrates effectiveness for knee flexion angle estimation, offering a practical tool for researchers and clinicians.