The proposed model in this paper is a collaborative multi-robot system that utilizes a combination of unmanned aerial vehicles (UAV), providing aerial fire detection surveillance, and unmanned ground vehicles (UGV), representing fire fighting vehicles. The mission of this team is to detect randomly scattered fires and extinguish them in a distributed manner. All UAVs can observe the environment and estimate the location of fires. However, each UAV specifically patrols a particular area to obtain more accurate information and therefore may have noisy or incomplete information about the other parts of the environment. After acquiring the fire locations each UAV creates an ordered list of targets based on the percentage of noise affecting their accuracy. Subsequently, these lists are transmitted to the UGVs based on a priority scheme. The UGV with the highest priority (in terms of computational power, sets of skills, communication bandwidth, etc.) receives the ordered list first and selects a target (fire location) by optimizing an objective function according to the targets’ order in the list and the shortest distance. However, if the UAVs are unable to contact the highest priority UGV, they will initiate a connection with the next UGV in the priority scheme. Once the UGV has selected a target, the remaining list is passed to the next highest priority UGV.
The strength of the developed algorithm is facilitating conflict resolution between the UGVs based on which they can autonomously find the most suitable location to attend. This allows the UGVs to make more informed decisions and prevents overlapping of resources. We have also implemented a practical testbed to investigate the performance of the proposed algorithm. The experimental results illustrate a high efficiency in performance of the multi-robot systems collaborating and distributing the task of firefighting amongst UGVs.