NFC-Based Map Switching for Efficient Robot Operation in Indoor Environments
Huichang Yun, Seungho Yoo · 2025
Currently, robots are rapidly developing for more complex and accurate tasks in various environments due to the development of AI. As the purpose of robot utilization deepens, robots need to understand their surroundings in more detail, and this is where 3D maps are used. However, 3D maps fundamentally require a lot of memory usage and computational power. This is fatal in environments with limited resources. Depending on the purpose of the robot's work, the robot might need a 3D map, or a 2D map may be sufficient. However, existing approaches that use a single 2D or 3D map lack flexibility to adapt to different situations, leading to inefficient use of memory and computational resources. In this paper, we solve these problems by dividing the map into map segments and using 2D and 3D map segments alternately through map switching. By independently managing the map segment, the proposed system enables easy updates and efficient overall map management. In addition, NFC tags placed at the map switching points between map segments, which can be a reference for map switching, and the automatic initial position estimation feature enables reliable initial position estimation. With these features, the proposed system optimizes memory and computational resources by using 2D maps for simple navigation tasks such as simple movement, and switching to 3D maps for complex tasks. We implement and evaluate the system using ROS 2 and Gazebo to demonstrate the operation and efficiency of the system.