Manipulation of Unknown Objects to Improve the Grasp Quality Using Tactile Information

Area of Science:

• Robotics
• Artificial Intelligence
• Mechanical Engineering

Background:

• Robotic manipulation of unknown objects is challenging due to complex geometric and mechanical properties.
• Existing methods often rely on extensive pre-programming or external sensing, limiting adaptability.

Purpose of the Study:

• To develop a novel, simple, and reactive approach for robotic manipulation of unknown objects.
• To improve grasp quality, hand configuration, and object positioning using only tactile feedback.
• To ensure object stability during manipulation without external sensors.

Main Methods:

• The approach utilizes tactile feedback for local contact information between fingertips and the object.
• Grasp optimization is performed online as a reactive procedure.
• Kinematic information is integrated with tactile data for manipulation control.

Main Results:

• The method successfully improves grasp quality, hand configuration, and object positioning.
• Object stability is maintained throughout the manipulation process.
• Experimental results demonstrate the efficiency and effectiveness of the proposed approach.

Conclusions:

• The novel approach enables effective manipulation of unknown objects using only tactile feedback.
• Online grasp optimization as a reactive procedure is a viable strategy for robotic manipulation.
• The method offers a simple yet efficient solution for improving robotic hand performance in grasping tasks.