Intelligent path planning of the unmanned aerial vehicle (UAV) swarm

Industrial technological developments have enabled the movement and maneuverability of Unmanned Aerial Vehicles (UAVs) to progress rapidly. Coordinating multiple UAVs can enable faster and more efficient mission execution compared to a single UAV. Especially in social areas such as entertainment, shipping, transportation and distribution; Its use in military fields such as surveillance, tracking and attack has accelerated research on swarm UAV systems. Swarm UAV systems contain control parameters sensitive to swarm and environment size, and physical definitions such as behavior-based, virtual structure, leader following or consensus may be needed in target-oriented missions. In the literature, algorithms based on these definitions used to ensure coordinated movement of swarm members can provide successful results in limited areas. However, swarm behavior is like social behavior that occurs automatically and not in a specific order. Therefore, in robust swarm topologies, the remaining members are expected to continue working together even if some are missing in dynamic environments. The main problem in this dynamic environment is that swarm parameters are flexible to suit the external environment. Additionally, not all members may need to perform the task in large swarms, and the swarm should be scalable with different groups of members. Therefore, to create an efficient swarm UAV topology that will work on other tasks in many different areas, a swarm system should be developed that encompasses the concepts of robustness, flexibility and scalability. This thesis has developed a swarm topology based on the Consensus-based Virtual Leader Tracking Algorithm (CVLTA), considering the deficiencies in the abovementioned issues, which will bring robustness, flexibility and scalability to system. This algorithm can successfully solve swarm sub-tasks such as aggregation/gathering, flight formation and formation control during flight, while the swarm tries to reach the target point by avoiding obstacles. In addition, with the applications, a swarm topology that is robust enough to continue its mission despite the loss of a random number of members at a random moment during the movement, scalable enough to create new flight formations, and flexible enough to provide formation control during flight has been revealed. Since the proposed algorithm performs instant (online) route planning, it may be relatively inefficient in large areas and regions containing large maps. Therefore, to provide a broader solution to the energy and cost problems in swarm UAV applications, the CVLTA algorithm was developed with GDRRT*-PSO, an intelligent path-planning algorithm. As a result, it can be claimed that the proposed methods can be superior to a significant part of the previous studies in the literature. Additionally, applications implemented in a simulation environment can run in real-time systems.