ROBOTIC ARM INTEGRATED WITH SMART CLEANER FOR OBJECT CLASSIFICATION USING YOLO

Abstract: The integration of computer vision with robotic arm manipulation significantly improves the efficiency, accuracy, and autonomy of modern robotic systems. This project focuses on the design and implementation of a robotic arm integrated with real-time object classification using the YOLO (You Only Look Once) algorithm. The system is developed to identify, classify, and manipulate objects accurately through a vision-guided mechanism, enabling intelligent pick-and-place operations in dynamic and unstructured environments.

A camera mounted above the workspace continuously captures live video streams. These frames are processed using a YOLO-based object detection model, selected for its high speed and single-stage detection architecture, which ensures real-time performance. The algorithm detects multiple objects simultaneously, providing bounding box coordinates, class labels, and confidence scores for each object. Based on the detected object’s position and category, control signals are generated and transmitted to the robotic arm controller for precise movement and gripping actions. The model is trained using a custom dataset to improve classification accuracy for specific target objects. Experimental results demonstrate reliable detection and successful manipulation under varying lighting conditions and object orientations. Overall, integrating YOLO with robotic arm control enhances system speed, adaptability, and operational accuracy compared to conventional vision-based robotic systems.

Keywords: Smart Cleaner, Object Classification, USB Camera, YOLO