Efficient Robot Identification in Multi-Robot Systems: A ROS-Based YOLOv8 Implementation for Real-World Scenarios

Enabling seamless interaction among multiple robots is essential for improving task execution in heterogeneous industrial environments, reducing human intervention while ensuring efficient manoeuvring and task processing. Accurately identifying different robots within the system is crucial for efficient task allocation, as it enables the assignment of appropriate tasks to the correct robotic unit, optimizing workflow and resource utilization. Efficient robot identification is required for perception and decision-making in multi-robot systems, enabling accurate recognition and classification for effective task allocation. This study presents an object detection framework designed to identify four classes of robots using YOLOv8 and ROS 2. A custom dataset was created, processed, and annotated to ensure high-quality training data through manual labelling. The training process utilizes YOLOv8, achieving 93.3% precision, 94.2% mAP, and 90.8% recall. ROS 2 Humble serves as middleware to enable smooth communication between nodes. The trained model is evaluated in both simulated and real-world environments. Experimental results demonstrated the system’s effectiveness in enhancing robotic vision tasks and improving object detection in real-time environments with multiple robots.