Targeted muscle effort distribution with exercise robots: Trajectory and resistance effects

Area of Science:

• Biomechanics
• Robotics
• Neuroscience

Background:

• Robotic exercise machines offer controlled rehabilitation and training.
• Understanding muscle effort distribution is key to optimizing these systems.
• Electromyography (EMG) measures muscle activity during exercise.

Purpose of the Study:

• To correlate muscle effort distribution with robotic exercise machine trajectory and resistance settings.
• To investigate the impact of elliptical trajectory orientation and stiffness on muscle activation patterns.
• To assess the predictive accuracy of an artificial neural network (ANN) for muscle effort over time.

Main Methods:

• Utilized a four degrees-of-freedom robot with impedance control for exercise protocols.
• Measured muscle effort distribution using electromyography (EMG) sensors.
• Trained an artificial neural network (ANN) on experimental data and evaluated its predictive performance.

Main Results:

• Identified relationships between robot trajectory (ellipse orientation) and resistance (stiffness) with muscle effort distribution.
• Demonstrated a decline in ANN model precision over time for predicting muscle effort.
• Observed that long-term muscle effort estimations are complex and may be influenced by fatigue.

Conclusions:

• Robotic exercise parameters significantly influence muscle activation patterns.
• The predictive accuracy of ANNs for muscle effort diminishes over time, suggesting dynamic changes in muscle response.
• Muscle fatigue likely contributes to time-varying dynamics, complicating long-term effort prediction in robotic rehabilitation.