Vision-Based Tactile Sensor Mechanism for the Estimation of Contact Position and Force Distribution Using Deep...

Area of Science:

• Robotics
• Computer Vision
• Sensor Technology

Background:

• Tactile sensing is crucial for robotic manipulation and interaction.
• Accurate estimation of contact position, area, and force distribution remains a challenge.
• Vision-based approaches offer a novel method for tactile sensing.

Purpose of the Study:

• To develop and validate a vision-based tactile sensor system.
• To train a neural network for estimating tactile contact parameters.
• To investigate the impact of sensor design on performance.

Main Methods:

• Utilized image-based information from a tactile sensor with input loads.
• Employed a stereo camera setup to capture sensor deformation.
• Modified a VGG-16 convolutional neural network for regression tasks using transfer learning.

Main Results:

• Successfully trained a neural network to estimate tactile contact position, area, and force distribution.
• Demonstrated the system's ability to handle various motions and input loads.
• Validated the model's performance with different tactile sensor designs (thickness, shape).

Conclusions:

• The developed vision-based tactile sensor system effectively estimates contact parameters.
• Neural network-based regression, enhanced by transfer learning, is suitable for this task.
• Sensor design considerations (thickness, material, shape) influence system accuracy.