Advanced IoT Gesture Recognition using Evolving Neural Networks
The project focuses on developing an Advanced Gesture Recognition System implemented on the ESP32 microcontroller.
This system leverages the power of Topology and Weight Evolving Artificial Neural Network (TWEANN) architecture. TWEANN is a highly effective evolutionary computation technique used to:
- Optimize Network Structure (Topology): Instead of using a fixed network design, TWEANN dynamically evolves the connections and number of neurons.
- Optimize Connection Weights (Weights): It simultaneously tunes the strength of the connections between neurons.
Key Enhancement and Rationale
The use of TWEANN is a crucial differentiator, as it allows the system to:
-Achieve High Accuracy with Minimal Resources: By generating a highly optimized, compact neural network structure specifically tailored for the gesture data, it ensures maximum recognition accuracy while meeting the tight memory and processing constraints of the low-power ESP32 embedded platform.
-Adapt to New Gestures: The evolutionary nature of TWEANN makes the system inherently adaptable and robust to variations in user movement and environmental changes.