Real-Time Hand Gesture Recognition in Clinical Settings: A Low-Power FMCW Radar Integrated Sensor System with...

Area of Science:

• * Biomedical Engineering
• * Signal Processing
• * Human-Computer Interaction

Background:

• * Contactless human-machine interaction (HMI) is crucial for integrated sensor systems in clinical settings.
• * Low-power solutions adaptable to edge computing are essential for efficient healthcare applications.
• * Existing radar sensing methods face challenges in robustness and computational efficiency for edge deployment.

Purpose of the Study:

• * To develop a real-time hand gesture recognition system using a low-power Frequency-Modulated Continuous Wave (FMCW) radar sensor.
• * To optimize the system for deployment on edge devices, addressing computational and power constraints.
• * To enhance robustness against environmental noise and complex clinical conditions.

Main Methods:

• * A novel Multiple Feature Fusion (MFF) framework integrating velocity profiles, angular variations, and spatial-temporal features.
• * A dual-stage processing architecture: adaptive energy thresholding for gesture segmentation and an attention-enhanced neural classifier.
• * Innovations include dynamic clutter suppression and multi-path cancellation for complex environments.

Main Results:

• * Achieved 98% detection recall and 93.87% classification accuracy via Leave-One-Subject-Out (LOSO) cross-validation.
• * Demonstrated high robustness against environmental noise and lower computational overhead compared to existing methods.
• * Real-time processing at 28 FPS on embedded hardware.

Conclusions:

• * The proposed low-power, edge-based FMCW radar system offers a robust and efficient solution for contactless HMI in healthcare.
• * Suitable for applications such as sterile medical control and patient monitoring, advancing radar sensing for edge computing.
• * Addresses critical efficiency and robustness challenges in clinical contactless interaction.