Efficient Traversability Mapping for Enhanced Robotic Firefighting Operations
Youngjoo. Choi, Hanwool Lee, Hwang Jung Hoon, Minji Park · 2024
This paper addresses the critical need for safer firefighting practices, highlighted by the increase in firefighter injuries and fatalities reported in the 2023 Fire Service Statistical Yearbook. With 239 injuries and three fatalities in 2022, the study underscores the urgency of integrating robotics into firefighting operations to enhance safety. Our focus is on rapidly generating accurate traversability maps. We employ advanced techniques, such as Inpaint and Smoothing Filters, to effectively process raw elevation data, ensuring continuous and reliable traversability calculations. Validation through LiDAR data from a Unitree Go1 robot in various environments demonstrates the effectiveness of our approach. Moreover, our Traversability Filter Model processes grid_map messages to produce probabilistic traversability values, offering improved granularity for training. Comparative results highlight the success of our method in enhancing traversability estimation, thereby optimizing robotic firefighting operations.