Development of an Accelerometer-based Data Acquisition System for Hand Gesture Recognition

This study presents a novel approach to Human-Robot Interaction (HRI) by employing gesture recognition to control robotic systems, such as bionic arms or manipulators, through accelerometer-based data collection. Using a low-cost glove equipped with an ESP32 microcontroller and accelerometer, data was gathered from 22 participants performing five distinct gestures: a clockwise circle, square, triangle, double tap, and the letter S. After filtering, these gestures were translated into features that trained a small neural network model designed for low-storage microcontrollers like the ESP32. With a performance accuracy of 90.4%, the neural network outperformed existing ensemble learning models, demonstrating its suitability for real-time gesture recognition tasks in embedded robotic systems. This highly efficient model enables seamless and intuitive control of robotic devices by facilitating responsive and natural HRI through gesture recognition. Designed specifically for compact and resource-limited environments, the model addresses the challenges associated with limited computational resources and restricted hardware capabilities. It provides a robust and scalable solution that can be easily integrated into various robotics applications, where gesture-based control is advantageous.