Cortex inspired model for inverse kinematics computation for a humanoid robotic finger.

Area of Science:

• Robotics
• Neuroscience
• Biomechanics

Background:

• Artificial anthropomorphic hands increasingly incorporate human biomechanical features to match human performance.
• Learning inverse kinematics for multi-jointed actuators with nonlinear sensory-motor relationships and joint coupling is challenging.

Purpose of the Study:

• To assess if a previously developed cortical model can learn the inverse kinematics of an actual anthropomorphic humanoid finger.
• To evaluate the model's ability to control a finger with coupled joints and pneumatic muscle actuation.

Main Methods:

• The study utilized a cortical model to learn the inverse kinematics of a physical anthropomorphic humanoid finger.
• The finger featured coupled joints and pneumatic muscle control, mimicking human biomechanics.

Main Results:

• The cortical model accurately and robustly performed single 3D reaching movements and complex motion patterns.
• The model's generated kinematics were comparable to human movement patterns.

Conclusions:

• The bioinspired cortical model successfully learned the complex inverse kinematics of the humanoid finger.
• This work advances the development of adaptive, robust, and flexible controllers for dexterous robotic and prosthetic hands.