Clustering-based task coordination to search and rescue teamwork of multiple agents

Abstract

It is important to have reasonable task coordination and path planning in rescue operations after a large-scale urban disaster. Whereas, there are many problems which can hamper rescue operations, such as communication obstacles, collapsed buildings, and secondary disaster. This article proposes a novel approach named ISODATA-K to achieve the task coordination and execution with heterogeneous ad hoc multi-agent. Inspired by the clustering analysis, ISODATA-K method, which does not require any input and threshold parameters, assigns the rescue agents to every area of the damaged city adaptively and efficiently. When the rescue agents get respective task, the path planning is done by A* algorithm which costs little time to find the relatively short route. The results of experiments demonstrate that the proposed method allows satisfactory rescue operations.