Development of Control Algorithms for an Adaptive Running Platform for a Musculoskeletal Rehabilitation System

Area of Science:

• Rehabilitation Engineering
• Biomechanics
• Human-Computer Interaction

Background:

• Adaptive treadmills are crucial for modern musculoskeletal rehabilitation, offering automated speed regulation based on user movement analysis.
• Current systems vary in control (manual vs. automatic) and hardware configuration, impacting user stability and training effectiveness.

Purpose of the Study:

• To experimentally compare control functions of adaptive treadmills for musculoskeletal rehabilitation.
• To assess the impact of hardware configuration and tracking systems on user stability and process smoothness.
• To determine optimal parameters for adaptive treadmill control functions.

Main Methods:

• Investigated two running platforms (varying length, with/without handrails) and two tracking systems (VR trackers, computer vision with MediaPipe Pose).
• Evaluated three control functions: linear, nonlinear, and proportional-integral-derivative (PID).
• Utilized a quality control criterion integrating metrics like positional stability, transient process duration/amplitude, and subjective user assessment.

Main Results:

• The linear control function with VR trackers and the PID function with computer vision (MediaPipe Pose) were identified as optimal.
• Established relationships between treadmill working area length, user comfort, and system performance.
• Formulated recommendations for treadmill platform length, tracking system selection, and control function parameters.

Conclusions:

• The findings provide practical guidance for developing adaptive rehabilitation simulators with improved control algorithms.
• Optimized control ensures smoother, more stable treadmill motion, enhancing user comfort, efficiency, and safety in rehabilitation.
• Recommendations support the design of advanced musculoskeletal rehabilitation systems.