A Multiagent-based Framework on RoboCup Simulation System for Enhancing Rescue Operation during Dynamic Disaster Environments

Disasters, such as fires, earthquakes, and hurricanes, inflict significant losses on human lives and economies. These unfortunate events have driven research into best-practice rescue methods via the RoboCupRescue competition and simulation system. One crucial aspect of rescue management is to ensure police clear ingresses and egresses to allow quick evacuations for citizens and efficient movement for other rescue personnel. To this end, existing rescue methods mainly rely on the A* path-planning method coupled with a nearest-neighbour clearance strategy. However, the A* path-planning method leads to routes selected for searching blockages that ignore the prevailing disaster conditions. The nearest-neighbour clearance strategy often results in curved clearance that takes more than the optimal times. Hence, to enhance the efficiency of police rescue, we propose a novel framework called adaptive A* graph routing with smoothed obstacle clearance. This framework incorporates an adaptive A* graph routing approach that will select paths based on the particular disaster conditions. Additionally, it uses a smoothed obstacle clearance strategy such that police will clear blocked roads in a particular ''smoothed'' direction. This framework optimizes route selection and obstacle clearance for police, improving rescue efficiency. Experimental results demonstrate that our proposed framework achieves superior rescue outcomes compared to the current baselines.