A novel solution approach for optimization of swarm UAV formation control problem under uncertain and dynamic constraints: Crazyfly 2.0 application

Abstract

In this article, Swarm UAV Formation Control Problem which is Homogeneous, Time-Window, Central Control Architecture, Uncertain, Dynamic Structure, and Leaderless (HTUDL-S-UAV) are obtained. For the solution of the problem, a new solution algorithm with generic and dynamic properties based on the annealing simulation algorithm is developed. The formation control architecture is determined as 5 layers and a new geometric approach to unit circle systematics has been developed to create formations. Tra-Coll and CCA algorithms are proposed in literature for tracking the trajectories of UAVs and preventing collisions. The application study is carried out on 9 different formations (Triangle, Square, Pentagon, V Shape, Crescent, Star, Rhombus Formation with 4 and 8, Line) and 7 different missions (Swarm Navigation Mission, Formation-Switching Mission, The mission to Remove UAV from the Swarm, The mission of adding UAV to the Swarm, Swarm Rotation Mission, Swarm Division/Unification Mission, Trajectory Tracking Mission as a Swarm) using Crazfly 2.0 drones in a ROS simulation program and on the real system. The proposed solution approach carried out all of these tasks in line with the specific conditions of the mission and achieved 3rd place in the Teknofest-2021 Swarm UAV Competition in Turkey. Furthermore, thanks to the study, it is aimed to create a decision support system for the optimal success of strategic and critically important Swarm UAV missions and other fleet autonomous systems.