Retrieving a file binder from a bookshelf using pseudo-curved trajectory generation for a foldable robotic hand

Area of Science:

• Robotics
• Human-Assistive Systems
• Mechatronics

Background:

• Human-assistive robots require precise trajectory control for manipulation in domestic environments.
• High-cost sensors increase robot construction expenses, limiting accessibility.
• Developing cost-effective manipulation methods is crucial for widespread adoption.

Purpose of the Study:

• To present a novel method for retrieving objects using a simple robotic hand system with a pseudo-curved trajectory.
• To reduce the cost of human-assistive robots by minimizing sensor requirements.
• To enhance the reliability and success rate of robotic grasping operations.

Main Methods:

• Designed and 3D-printed a simple claw mechanism with a switch sensor for object grasping.
• Integrated the claw mechanism with a 6-DOF foldable robotic hand.
• Developed a pseudo-curved trajectory generation method utilizing the switch sensor data.
• Implemented a control system for failure recovery during the grasping task.
• Added a rubber sponge to the claw for improved flexibility and success rate.

Main Results:

• Successfully retrieved a file binder from a bookshelf using the developed system and pseudo-curved trajectory.
• Identified manipulator link rotational speed and claw vibration as key factors affecting success rate.
• Demonstrated significant improvement in success rate after adding a rubber sponge to the claw mechanism.
• Validated the effectiveness of the failure recovery control system through experiments.

Conclusions:

• The proposed method enables cost-effective object retrieval for human-assistive robots.
• Pseudo-curved trajectory generation with simple sensors is a viable approach for robotic manipulation.
• System modifications, such as adding flexibility and failure recovery, are critical for robust performance.
• This research contributes to the development of more affordable and reliable assistive robotic technologies.