Robust identification of multi-joint human arm impedance based on dynamics decomposition: a modeling study

Area of Science:

• Robotics
• Biomechanics
• Control Systems

Background:

• Accurate impedance estimation is crucial for human-robot interaction and understanding limb dynamics.
• Traditional methods struggle with the complexity of multi-joint systems, leading to unreliable identification.
• Increasing degrees of freedom (DOF) in human arms and robotic exoskeletons presents significant estimation challenges.

Purpose of the Study:

• To develop a robust, unbiased, and tractable method for multi-joint human arm impedance estimation.
• To address the intractability of identifying complex multi-input multi-output (MIMO) systems.
• To provide a systematic approach for dynamics decomposition in robotic and biomechanical systems.

Main Methods:

• A novel dynamics decomposition technique was employed, breaking down MIMO systems into single-input multi-output (SIMO) subsystems.
• The method was validated using simulations of a human arm model and a 2-DOF exoskeleton robot.
• Simulations incorporated varying levels of sensor resolution and nonlinear friction to test robustness.

Main Results:

• The proposed dynamics decomposition method demonstrated accurate and robust impedance estimation for multi-joint systems.
• The approach proved effective even under challenging conditions, including sensor noise and nonlinear friction.
• Validation confirmed the method's reliability in identifying the dynamics of both biological and artificial limbs.

Conclusions:

• The developed dynamics decomposition method offers a reliable solution for multi-joint impedance estimation.
• This approach enhances the tractability and accuracy of identifying complex robotic and biomechanical systems.
• The method is adaptable for identifying sophisticated systems with a higher number of joints and degrees of freedom.