Publication

Abstract : In the ever-changing environments of robotics and artificial intelligence, the autonomous recovery of objects remains paramount, especially in indoor settings comprising extensively variable conditions. Fetch Mate is an AI-based autonomous system for the interior spaces, mapping by itself, detecting objects, and returning lost items upon user request. The system uses LiDAR-based Heuristic Simultaneous Localization And Mapping (Hector SLAM) for accurate navigation and ResNet object detection. Improving localization accuracy is accomplished using ORB. ROS2 is being adopted as the middleware for fast data exchange between different software components, and MQTT is being employed for mobile app interaction in real-time. The robot is based on a four-wheel drive with a Raspberry Pi, LiDAR, ultrasonic sensor, camera module, and DC motors. Performance measures are mapping accuracy, object detection accuracy, and retrieval rate. Application areas include home care, office automation, and assistance for disabled persons.