A Multiagent Mission Coordination System for Continuous Situational Awareness of Bushfires

Abstract

This paper devises a multi-agent Mission Coordinating Architecture (MCA) to achieve continuous situational awareness (SA) in bushfires, which can help with quick detection and accurate response to the hazards. In this paper, we use Unmanned Aerial Vehicles (UAVs) as instantiations of physical agents. MCA is a scalable architecture and aims to provide the UAVs with parallel mission plans, adopting a fire spread probability map and the Fuzzy C-means method to avoid mission overlap and duplicate information. The architecture is then enhanced by integrating a synchronized communication framework to facilitate UAVs’ adaptive cooperation and total flight time optimization. Furthermore, integrating the communication framework minimizes the number of deployed UAVs to fully cover the same area, saving considerable cost and energy compared to the Parallel Mode. The scalability challenge of the Parallel Mode, determining the required number of UAVs to cover the entire area, and the efficiency of the mission planning algorithm are thoroughly investigated and compared to the performance of the Communication Mode. Finally, the simulation results prove the MCA’s effectiveness in enhancing the UAVs’ exploration capability, resulting in comprehensive monitoring of all affected areas. Note to Practitioners—The proposed architecture in this research could offer flexibility to increase the number of UAVs on demand without requiring a change and adjustment of the parameters. Furthermore, a synchronized communication framework enhances MCA, enabling all UAVs to share their resources and exploit residual battery time to assist each other, manage the overall operation time, and reduce the total operation cost.