Autonomous Navigation with Active SLAM for Disinfecting Robot

Abstract

This paper introduces autonomous navigation using the active SLAM for a disinfecting robot. For a safe service and navigation of the disinfecting robot, sensors such as distance sensors and four cameras were set up. To prevent getting stuck between columns and obstacles, in this paper, we generate fake distance data from a LiDAR sensor. We also propose a planner that uses the active SLAM to enable SLAM and path planning simultaneously to generate coordinates that allow the robot to navigate to the goal position. For safe sterilization using UV-C, furthermore, human detection is very crucial because UV-C radiation can be fatal to humans. Therefore, a MobileNetSSD was used to detect humans accurately with 15 FPS for the disinfection robot. Using these approaches, we present autonomous navigation with the A* and DWA algorithms for disinfection. Finally, through experiments, we verified our system of autonomous navigation for the disinfecting robot in a simulation and a real environment. In particular, we confirmed that the disinfecting robot can disinfect effectively through the possibility of sterilization by UVC dosimeters.