dead reckoning on the Össur Power Knee

towards pedestrian dead reckoning and incline estimation on the Össur Power Knee

class: Quantifying Human Motion Through Wearable Sensors
term: winter 2025
write up: forthcoming following publication

This study explores the feasibility of using the limited sensor suite of the Ossur Power Knee to estimate gait speed and incline in a pilot study. We use a Kalman filter approach for fusing accelerometer and gyroscope signals from the shank- mounted inertial measurement unit (IMU) to estimate sensor orientation and transform IMU accelerations into the global frame. Stride segmentation is performed using thresholds on Power Knee’s strain gauge signals to identify gait events. We then utilize pedestrian dead reckoning techniques to estimate shank position over time. We remove gravity from the global frame acceleration, use trapezoid integration to estimate velocity and position, and apply zero-velocity updates per stride to minimize drift. From the reconstructed motion, we derive estimates of walking speed and terrain slope. Although this work is limited to a single participant and short walking bouts, preliminary results suggest that meaningful task-level insights can be estimated even with the limited sensor suite on the Power Knee.