Noncontact Electromagnetic Wireless Recognition for Prosthesis Based on Intelligent Metasurface.

Area of Science:

• Electromagnetics
• Metasurface Technology
• Human-Machine Interaction

Background:

• Hand gesture recognition is crucial for advancing human-machine interaction (HMI).
• Existing methods often lack non-contact capabilities or sufficient accuracy.
• Developments in artificial intelligence (AI) and the Internet of Things (IoT) necessitate improved HMI solutions.

Purpose of the Study:

• To propose a non-contact hand gesture recognition method using an intelligent metasurface.
• To leverage dynamic electromagnetic (EM) focusing for comprehensive echo data acquisition.
• To enhance the accuracy of gesture recognition for HMI applications.

Main Methods:

• A transmissive programmable metasurface was designed to control EM focusing.
• The metasurface illuminated the forearm, capturing detailed echo data.
• Machine learning, specifically a support vector machine algorithm, was employed for gesture classification.
• Linear discriminant analysis and Fisher score were used to analyze focusing spot influences.

Main Results:

• The programmable metasurface acquired unique echo coefficient variations from muscle perturbations.
• Gesture recognition accuracy was higher using the metasurface compared to traditional passive antennas.
• The system demonstrated successful classification of hand gestures based on focused EM data.

Conclusions:

• The proposed metasurface-based non-contact wireless design offers a high-accuracy solution for hand gesture recognition.
• This technology provides a viable pathway for advanced human-machine interaction.
• Intelligent metasurfaces enable enhanced data acquisition for sophisticated AI-driven applications.