Electrostatic Enhanced Dual-Mode Electronic Skin for Multifunctional Robotic Hands Capable of Object Shape and...

Area of Science:

• Robotics
• Materials Science
• Artificial Intelligence

Background:

• Multimodal perception is crucial for robots operating in dynamic, unstructured environments.
• Existing robotic sensing methods often lack the versatility for comprehensive object interaction.
• Intelligent interaction requires robots to perceive both physical form and material characteristics.

Purpose of the Study:

• To develop a multifunctional robotic hand with advanced object recognition capabilities.
• To introduce a novel dual-mode electronic skin (e-skin) for integrated non-contact and contact sensing.
• To enhance robotic systems' ability to interact intelligently in diverse environments.

Main Methods:

• Fabrication of a dual-mode e-skin using a polarized expanded polytetrafluoroethylene electret embedded in Ecoflex.
• Implementation of non-contact sensing via the electrostatic-field effect and contact sensing via the triboelectric effect.
• Integration of the e-skin with a robotic arm and utilization of a long short-term memory neural network for data analysis.

Main Results:

• The dual-mode e-skin demonstrated enhanced non-contact sensing range and intensity due to its embedded electret architecture.
• The robotic hand achieved high accuracies: 100% for object-shape recognition and 97.35% for object-material recognition.
• The system successfully integrated multimodal sensing for intelligent object identification.

Conclusions:

• The proposed dual-mode e-skin offers a versatile multimodal sensing interface for robotic platforms.
• This technology significantly advances autonomous and intelligent robotic interactions by improving object perception.
• The e-skin's unique design overcomes limitations of conventional sensing approaches.