Blind Manipulation of Deformable Objects Based on Force Sensing and Finite Element Modeling.

Area of Science:

• Robotics
• Mechanical Engineering
• Computational Mechanics

Background:

• Accurate estimation and manipulation of object deformation are crucial for advanced robotic applications.
• Traditional methods often rely on vision-based sensing, which can be limited in certain environments.
• Developing vision-free methods for deformation control is an active research area.

Purpose of the Study:

• To present a novel pipeline for simultaneous estimation and manipulation of object deformation.
• To demonstrate the feasibility of using only force sensing and a Finite Element Method (FEM) model for this task.
• To evaluate the accuracy and effectiveness of the proposed approach in real-world experiments.

Main Methods:

• A pipeline integrating a sensor model, a deformation model, and a pose controller was developed.
• The sensor model estimates contact forces from sensor data.
• The deformation model updates the object's volumetric mesh based on estimated forces, and the pose controller manipulates the object to achieve a desired pose.

Main Results:

• The proposed pipeline accurately estimates and manipulates object deformations using only force sensing.
• Experimental results with a robot manipulator and force-torque sensor validate the system's performance.
• The approach successfully operates without the need for vision sensors.

Conclusions:

• The developed pipeline offers a robust, vision-free solution for simultaneous deformation estimation and manipulation.
• This method enhances the capabilities of robotic systems in tasks requiring precise interaction with deformable objects.
• The findings open new avenues for tactile-based robotic manipulation and control.