Gesture Controlled Robotic Hand Designed for Enhancing Industrial Automation and Innovation

Abstract

The robotic hand was controlled by hand gestures using a camera to track hand movement and replicate those motions. No wearable sensors or physical controllers were required for this research. The system was built using cost effective materials, which made it cheaper to produce. Simple components were used, which made the system easier to repair. It was developed to reduce direct human involvement in hazardous environments. At the centre of this work, computer vision was applied. Gestures were captured through MediaPipe, a pretrained machine learning model. It was used to detect 21 hand landmarks and calculate finger angles. In addition, a wrist mode was implemented to control the motion of the hand itself. This allowed a greater range of mobility. The core principle of the gesture controlled robotic hand was to examine whether an artificial intelligence integrated system could support industrial tasks in a safe and affordable manner while promoting resilient infrastructure, sustainable industrialization, and responsible technological innovation aligned with SDG 9 goals.