Target tracking optimization with drone swarm

In recent years, the use of drones, which are in the class of unmanned aerial vehicles, for military, civil (hobby and commercial) and academic purposes has become increasingly common. It is mostly used in reconnaissance, search, rescue and attack missions in the military field, cargo transportation in civil use, imaging, mapping and hobby activities, and in the academic field for the development of flight management and control software. One of the most important problems of using military drones is the limitation of the distance between the Drone and the command center. In order to overcome this distance limitation, some studies have been carried out on autonomous drones that move to the desired coordinates and successful results have been obtained as a result of these studies. However, when autonomous drones, which are used alone for military purposes, are faced with certain problems, they cannot provide the desired data transfer to the command center and cannot successfully perform the task they need to perform. Studies are carried out on herd algorithms to overcome this distance problem of drones. It is also very popular to work on topics such as multiple Drone interaction and task sharing. Researchers are focusing on new algorithms that enable drones to move in swarm, especially for reconnaissance, search, rescue and attack missions in the military field. In this thesis, an objective function has been defined for the most appropriate tracking of a moving target with the drone swarm and target tracking optimization algorithms have been developed that optimize according to this goal function. In the study, target tracking and destruction missions defined for the drone swarm were tested in Unity simulation environment with three different optimization algorithms. In the study, Particle Swarm Optimization (PSO), which drives drone swarm, Artificial Bee Colony Algorithm (YAK), which reveals the behaviors of bees by acting together in nature, and Ant Colony Algorithm (KKA), which is revealed by examining the behavior of ants in nature, are discussed. Even if one of the Drones in the herd is damaged during the defined mission, other Drones can continue their duties by taking the most appropriate position according to the algorithm structure used. In the simulation study, the sample task of the drone herd getting up from a station and following a vehicle, positioning it on the vehicle in a circular orbit and destroying the vehicle was defined. For this purpose, a target vehicle and a ground station where the desired number of Drones can take off were designed in the Unity simulation environment. Assuming that the location information of the vehicle is received via satellite, the position of the drones with respect to the target vehicle has been provided by the determined optimization algorithms. In the study, the performances of the optimization algorithms were compared based on the objective function defined for the task.