Automated Maneuvering Decision for UAVs in Forest Surveillance and Fire Detection Missions

As a exible, efficient, and powerful platform for a variety of practical applications, the unmanned aerial vehicle (UAV) has attracted increasing attention in the field of forest re monitoring, detection, and tracking in recent years. Generally, more complicated and heavier tasks can be achieved by using a team of UAVs other than a single UAV. However, a possible problem is that faults may occur on a UAV in the formation or the power may gradually run out during the task, so that the disabled UAV must be replaced by a newly assigned UAV from the base. Hence, the problem becomes how to guide the new UAV to join the team in an optimal manner, and then to maintain the formation during the remaining task. Meanwhile, collision avoidance should be involved to avoid hazardous impacts from other formation members during the whole mission. To solve these challenges, an automated maneuvering decision strategy consisting of a tailored scoring function and a game tree decision algorithm is developed in this study. The newly assigned UAV can be guided to join the patrol team from arbitrary initial locations, and the formation can be maintained during the patrol task in the presence of disturbances. Numerical simulations are conducted to validate the proposed decision strategy.