Çok amaçlı insansız hava aracı rotalama

Unmanned Aerial Vehicles (UAVs) are special types of aircraft operating without a human pilot on board. In this thesis, the multi-objective UAV route planning is addressed by studying three different problems. In the first study, the bi-objective route planning problem of a UAV that is tasked with visiting all targets located in an enemy region monitored by radars is addressed. The aim is to determine both the visiting order of the targets and the specific trajectories to use between consecutive target pairs that minimize the total flight duration and risk of being detected. The terrain the UAV moves is considered as continuous terrain allowing infinitely many trajectory options between targets. We develop a generic preference-based multi-objective evolutionary algorithm (MOEA) that can be used for any multi-objective optimization problem, and adapt it to the UAV route planning problem. In the second study, the first study is extended by considering an additional objective that is maximizing the collected information. We address this problem adapting the MOEA developed in the first study. In both studies, the results show that the developed algorithm is able to converge to preferred regions of the Pareto-optimal frontier, provide a general idea on the structure and positioning of the entire Pareto-optimal frontier, and adapt to the changes in preferences quickly. In the third study, the UAV route planning problem is considered as a multi-objective, multi-connection orienteering problem with time dependent prizes. A route plan involves the decisions of which targets to visit, the order of visit to the selected targets, and the trajectories to follow between consecutively-visited targets. The objectives are maximizing the collected information from the operation area while minimizing the mission time and the radar detection threat on the entire route. We discretize the continuous movement space of the UAV by representing infinitely many efficient trajectories between target pairs with a finite trajectory set. We first formulate a mixed integer programming (MIP) model and then develop a hybrid algorithm to decrease the computational requirements of the model. The hybrid algorithm consists of a heuristic route search algorithm that approximates the optimal solution and a more manageable MIP model that finds some bounds for the optimal objective function value. The results show that the developed algorithm is able to reduce the computational requirements of the exact model. A demonstration on the Colorado state of the U.S. is performed as a case study.