Markerless motion tracking to quantify behavioral changes during robot-assisted gait training: A validation study

Abstract

Introduction: Measuring kinematic behavior during robot-assisted gait therapy requires either laborious set up of a marker-based motion capture system or relies on the internal sensors of devices that may not cover all relevant degrees of freedom. This presents a major barrier for the adoption of kinematic measurements in the normal clinical schedule. However, to advance the field of robot-assisted therapy many insights could be gained from evaluating patient behavior during regular therapies.Methods: For this reason, we recently developed and validated a method for extracting kinematics from recordings of a low-cost RGB-D sensor, which relies on a virtual 3D body model to estimate the patient’s body shape and pose in each frame. The present study aimed to evaluate the robustness of the method to the presence of a lower limb exoskeleton. 10 healthy children without gait impairment walked on a treadmill with and without wearing the exoskeleton to evaluate the estimated body shape, and 8 custom stickers were placed on the body to evaluate the accuracy of estimated poses.Results & Conclusion: We found that the shape is generally robust to wearing the exoskeleton, and systematic pose tracking errors were around 5 mm. Therefore, the method can be a valuable measurement tool for the clinical evaluation, e.g., to measure compensatory movements of the trunk.

Clustering trunk movements of children and adolescents with neurological gait disorders undergoing robot-assisted gait therapy: the functional ability determines if actuated pelvis movements are clinically useful

“…They were recorded at 30 Hz with an Azure Kinect DK RGB-D camera (Microsoft, Seattle, USA) placed directly in front of the patients to minimize occlusions by the Lokomat. The recordings were transformed offline into kinematic data using a custom 3D body tracking method, which fits a virtual body model [ 20 ] to the recorded point clouds of the Azure Kinect DK [ 21 ]. The model provides kinematic data as three-dimensional time series for 24 joint positions and angles.…”

“…To reduce carry-over effects between the conditions, only the mean of the last 25 strides of each condition was included in the analysis. In addition, reference data from 10 typically developing children and adolescents during treadmill walking were used from a different study [ 21 ] and processed with the same approach.

Fig.

…”

The FreeD module’s lateral translation timing in the gait robot Lokomat: a manual adaptation is necessary

“…The frontal video recordings of the gait pattern were processed offline with a method previously validated by van Dellen et al [ 12 ]. In short, RGB-D data were transformed to three-dimensional point clouds.…”

1D-Convolutional Neural Networks can Quantify Therapy Content of Children and Adolescents Walking in a Robot-Assisted Gait Trainer

“…The electrodes were placed according to the SENIAM guidelines [14] and slightly shifted if the designated position was covered by the straps of the Lokomat. In addition, the trials were recorded with an RGB-D camera (Azure Kinect, Microsoft, United States) to identify heel strikes [15].…”