An anthropomorphic soft skeleton hand exploiting conditional models for piano playing.

Area of Science:

• Robotics and Mechanical Engineering
• Soft Robotics and Biomechanics

Background:

• Achieving human-like dexterity and adaptability in robotic manipulators remains a significant research challenge.
• Existing robotic hands often lack the nuanced control and behavioral diversity seen in biological systems.

Purpose of the Study:

• To explore the use of passive dynamics in mechanically complex systems to enhance robotic manipulator adaptability and dexterity.
• To introduce and validate a "conditional model" approach for exploiting anisotropic stiffness in soft-rigid hybrid robotic hands.

Main Methods:

• Utilized state-of-the-art multimaterial 3D printing to design and construct an anthropomorphic skeleton hand with anisotropic mechanical stiffness.
• Developed and applied the "conditional model" concept, linking physical configuration, environmental conditions, and conditional actuation.
• Leveraged passivity-based dynamic interactions for joint actuation through conditional models.

Main Results:

• The developed skeleton hand demonstrated anisotropic mechanical stiffness, enabling complex passive structures.
• The conditional model approach facilitated altered physical configurations and actuation, allowing a single hand to perform diverse piano music phrases.
• Significantly improved dynamic behaviors and interactions with the piano were observed compared to rigid end effectors.

Conclusions:

• Passive dynamics and anisotropic stiffness are viable strategies for creating adaptable and dexterous robotic hands.
• The "conditional model" offers a novel framework for controlling complex soft-rigid hybrid robotic systems.
• This approach holds promise for advancing robotic manipulation in tasks requiring fine motor skills and environmental interaction.