AI Vision Powered Robotic Arm

Abstract

Robotics and artificial intelligence together represent one of the most transformative frontiers in modern technology. This paper presents the design and implementation of an AI Vision Powered Robotic Arm that integrates real-time object detection, spatial analysis, and precise servo-controlled manipulation. The system employs an ESP32-CAM module as an edge vision device, streaming live MJPEG video over a local Wi-Fi network at approximately 12 frames per second. A Python-based AI backend processes this video stream using the YOLOv8 object detection framework to identify, classify, and localize target objects in real time. Once a target is detected, the system computes the object coordinates, estimates the spatial position, and translates this information into servo motor commands delivered to an Arduino Uno controller via serial communication at 9600 baud.

The Arduino Uno governs a four-degree-of-freedom robotic arm comprising Base, Shoulder, Elbow, and Gripper servos, with smooth motion interpolation and an HC-SR04 ultrasonic proximity sensor for real-time distance measurement.

The system demonstrates successful real-time object detection at 12 FPS, reliable serial communication between the AI module and the robotic controller, and repeatable pick-and-place operations with a 92 percent success rate across 50 controlled trials.