Abstract

We seek to use dimensionality reduction to simplify the difficult task of controlling a lower limb prosthesis. Though many techniques for dimensionality reduction have been described, it is not clear which is the most appropriate for human gait data. In this study, we first compare how Principal Component Analysis (PCA) and an autoencoder on poses (Pose-AE) transform human kinematics data during flat ground and stair walking. Second, we compare the performance of PCA, Pose-AE and a new autoencoder trained on full human movement trajectories (Move-AE) in order to capture the time varying properties of gait. We compare these methods for both movement classification and identifying the individual. These are key capabilities for identifying useful data representations for prosthetic control. We first find that Pose-AE outperforms PCA on dimensionality reduction by achieving a higher Variance Accounted For (VAF) across flat ground walking data, stairs data, and undirected natural movements. We then find in our second task that Move-AE significantly outperforms both PCA and Pose-AE on movement classification and individual identification tasks. This suggests the autoencoder is more suitable than PCA for dimensionality reduction of human gait, and can be used to encode useful representations of entire movements to facilitate prosthetic control tasks.

Highlights

  • Models of human gait are the foundation upon which lower limb prosthesis controllers are built

  • We have previously explored how Principal Component Analysis (PCA) compares to an autoencoder for dimensionality reduction of hand kinematics, as it pertains to priorities for prosthetic control Portnova-Fahreeva et al (2020)

  • We have previously demonstrated the viability of using machine learning to predict joint kinematics for lower limb prosthesis control Rai et al (2020) Rai and Rombokas (2019)

Read more

Summary

Introduction

Models of human gait are the foundation upon which lower limb prosthesis controllers are built. Variable damping knees use on-board sensors to detect speed and phase, adjusting knee and ankle joint control parameters to mimic human gait Highsmith et al (2010). Today, powered prostheses that generate work during gait are gaining in popularity in research circles Azocar et al (2020). The challenge of controlling prostheses has been recently brought again to attention Iandolo et al (2019) Tucker et al (2015), and only highlighted by the untapped potential of powered devices to restore mobility. We assert that generating useful representations of human movement is necessary to unlock the potential of such devices

Objectives
Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call