Dimensionality Reduction of Human Gait for Prosthetic Control

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.

Deep learning with an attention mechanism for continuous biomechanical motion estimation across varied activities

“…The selection of input features also is a time-consuming process, and some components of the IMU signals might be found to make little contribution to the prediction precision ( Renani et al, 2021 ). Before the rise of deep learning, some feature engineering methods such as Principal Component Analysis (PCA) remain popular to pre-processing gait data, and then the transformed data were inputted into a regressor/classifier based on a traditional machine learning algorithm such as a Support Vector Machine (SVM) or Decision Tree (DT) ( Boe et al, 2021 ; Lotfi and Kedir-Talha, 2022 ). However, all measurable IMU physical quantities were directly fed into the neural network models in this work, and we rely on the automatic extraction of beneficial intermediate features and abandonment of useless components.…”

Unsupervised Machine Learning on Motion Capture Data Uncovers Movement Strategies in Low Back Pain

“…For the analysis, the biomechanics data were averaged across trials and repetitions of sit-to-stand as PCA suffers from intra-subject correlation and assumes independence of observations. Several studies on the kinematics of human movement comparing linear and nonlinear analytical methodologies have demonstrated that nonlinear methods outperform linear in the percent variance accounted for in the data ( Harbourne et al, 2009 ; Portnova-Fahreeva et al, 2020 ; Boe et al, 2021 ). Furthermore, if the variables are linearly related, then PCA and NLPCA both result in the same solution.…”

Deep learning and session-specific rapid recalibration for dynamic hand gesture recognition from EMG

“…However, apart from the output layer, the symmetric form Hinton and Salakhutdinov (2006) of the autoencoder is used, where the decoder consists of a mirror network to the encoder. This encoder has previously been shown proven to extract useful latent representations for lower-limb activities Boe et al (2021) and hand movements Portnova-Fahreeva et al (2020 , 2022) .…”