AI-Driven Smart Glove

Overview

The AI-Driven Smart Glove is an innovative project designed to capture joint motion and tactile feedback using a 3D-printed piezoelectric sensor. The glove uses a combination of triboelectric sensing, signal processing, and advanced hybrid LSTM-Transformer models to accurately detect and classify finger movements. This project integrates Arduino for real-time data collection and AI for enhanced motion detection.

Project Components

1. 3D-Printed PVDF Triboelectric Sensor

• A Polyvinylidene Fluoride (PVDF)-based triboelectric sensor is designed to detect joint movement and provide tactile feedback.
• The sensor is integrated into the glove for adaptive motion sensing, capturing real-time finger gestures.

2. Arduino Program

• An Arduino program is used to interface with the piezoelectric sensor, allowing it to capture data from the sensor and send it to a processing unit for real-time motion detection.
• The program reads the voltage generated by finger movements, which is then processed to detect specific gestures.

3. Signal Processing

• Signal processing techniques are applied to clean and prepare the raw data collected from the sensor. This includes noise reduction and normalization to improve the accuracy of motion detection.

4. Hybrid LSTM-Transformer Model

• The Hybrid LSTM-Transformer model is applied to the processed sensor data for accurate motion classification.
• LSTM (Long Short-Term Memory) layers capture sequential dependencies in the motion data, while Transformer layers enhance performance by modeling long-range dependencies.
• This hybrid approach significantly improves the glove’s ability to classify finger movements accurately.

Features

• Real-time Motion Detection: Arduino processes sensor data for immediate response.
• AI-Powered Gesture Classification: Uses machine learning models to identify various hand gestures.
• Customizable Design: The 3D-printed sensor and glove can be modified for different user needs.
• Tactile Feedback: Provides feedback based on the detected movements, improving user interaction.

Setup and Installation

Hardware Requirements

• 3D-printed PVDF triboelectric sensor
• Arduino board (e.g., Arduino Uno)
• Piezoelectric sensor
• USB cable for Arduino connection
• Smartphone or PC for model deployment and testing (optional)

Software Requirements

• Arduino IDE for programming the Arduino board.
• Python for signal processing and training the AI model.
• Libraries:
  • numpy
  • tensorflow (for LSTM-Transformer model)
  • matplotlib (for visualization)
  • scikit-learn (for data preprocessing)