Adapting Dynamic Programming-Based Floyd-Warshall Algorithm for Escape Routes in Smoke-Filled Disaster Environments

초록

Recent fires and disasters in large-scale complex spaces often result in massive casualties because dense smoke and low visibility make rapid evacuation difficult. To address this issue, this paper proposes an optimized evacuation-route planning system based on multimodal sensor fusion and a dynamic programming-based Floyd-Warshall algorithm. The proposed system integrates sensors such as vision, LiDAR, thermal cameras, and UWB radar to perform robust situation awareness and localization even in smoky environments. The indoor disaster space is modeled as a weighted graph composed of nodes, edges, exits, obstacles, and hazardous areas. The Floyd-Warshall algorithm is then applied to compute shortest and safest paths among all node pairs by updating route costs through dynamic programming. Experimental results from smoke-filled fire scenarios demonstrate that the proposed method provides stable path planning, rapid route retrieval after cost-table generation, and reliable alternative path selection when environmental risk weights are updated. The proposed approach is expected to improve the practicality and safety of autonomous evacuation guide robots in disaster environments.

키워드