A 4-layer printed circuit wristband using tomography to map electrical impedance for real-time gesture recognition via SVM.
Overview
The Tomographic Gesture Recognition Wristband is a novel wearable device that detects hand gestures in real-time using electrical impedance tomography (EIT). By measuring the electrical properties of the wrist tissue, the device can infer muscle and tendon positions to classify gestures.
Technical Architecture
Hardware Design
- PCB: Custom 4-layer printed circuit board designed from scratch
- Electrodes: Array of electrodes for multi-path impedance measurements
- Microcontroller: Low-power ARM processor for signal processing and wireless communication
- Power: Rechargeable LiPo battery with extended runtime
Software & Machine Learning
- Algorithm: Support Vector Machine (SVM) classifier for gesture recognition
- Training Data: Collected from multiple users performing standardized gestures
- Feature Extraction: Time-domain and frequency-domain features from impedance signals
- Accuracy: Achieved reliable classification across 6 distinct gestures
Development Timeline
This was an independent project spanning from December 2021 to August 2023, involving complete end-to-end development:
- Research Phase: Literature review of EIT principles and wearable sensing
- Hardware Design: Schematic capture, PCB layout, and component selection
- Fabrication: PCB manufacturing and assembly
- Software Development: Firmware, signal processing, and ML model training
- Testing & Iteration: User studies and algorithm refinement
Applications
Potential applications include hands-free device control, accessibility tools for users with limited mobility, and gaming interfaces.