Abstract / Summary
Abstract Indoor building fires create rapidly changing heat, smoke, toxic-gas and visibility conditions, making emergency route selection difficult when only geometric accessibility is considered. Small unmanned aerial vehicles (UAVs) can support pre-entry fire-response reconnaissance by inspecting smoke-affected corridors, potential access routes and hazardous rooms while reducing direct firefighter exposure. However, many routing studies still represent fire as a point source, a prescribed hazard zone or a single static risk map. This study proposes a fire-risk field enrichment framework that converts transient Fire Dynamics Simulator (FDS)/PyroSim outputs into route-level risk information for indoor UAV emergency routing. Temperature, carbon monoxide (CO) concentration and visibility fields are sampled at UAV flight height, normalized into risk channels and fused into time-indexed fire-risk fields. These fields are combined with a hard obstacle occupancy map and embedded into a cost-augmented $A^*$ planner. Robot Operating System 2 (ROS2)-Gazebo set-pose replay is used as an intermediate robotics-side check of spatial consistency and obstacle feasibility. For a fixed start--goal task, the proposed planner reduced cumulative sampled risk by 99.67\% compared with Standard $A^*$. Across five late-stage start--goal cases, reductions ranged from 37.9\% to 99.9\%. Offline dynamic risk-map replanning further reduced sampled risk by 28.32\%. The results show that transient fire simulation can provide actionable route-level support for indoor UAV fire-response reconnaissance.