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

Area of Science:

• Robotics
• Biomechanics
• Rehabilitation Engineering

Background:

• Clinical gait analysis for robot-assisted therapy is hindered by complex motion capture setups.
• Internal device sensors often lack comprehensive kinematic data.
• Accurate kinematic measurements are crucial for advancing robotic gait rehabilitation.

Purpose of the Study:

• To evaluate the robustness of an RGB-D sensor-based kinematic measurement method when a lower limb exoskeleton is worn.
• To assess the accuracy of estimated body shape and pose during treadmill walking with exoskeleton use.
• To determine the feasibility of this method for clinical gait evaluation.

Main Methods:

• Developed and validated a method using a low-cost RGB-D sensor and a virtual 3D body model.
• 10 healthy children walked on a treadmill with and without a lower limb exoskeleton.
• Custom stickers were used to evaluate pose tracking accuracy, with body shape assessed simultaneously.

Main Results:

• The RGB-D sensor method demonstrated robustness in estimating body shape with exoskeleton use.
• Systematic pose tracking errors were found to be approximately 5 mm.
• The findings support the method's reliability in clinical settings.

Conclusions:

• The RGB-D sensor-based kinematic analysis method is robust and accurate for evaluating gait during robot-assisted therapy, even with exoskeletons.
• This non-invasive technique can be a valuable tool for measuring compensatory movements in clinical settings.
• The method facilitates easier adoption into standard clinical schedules for gait analysis.