| Abstract [eng] |
This Master's thesis presents a study of UAV swarm possibilities. Different tools such as Robot Operating System framework, GAZEBO and Rviz simulator has been used in order to test and analyze different UAV swarm topologies in simulated environment. Mapping technique used to map the environment simultaneously during flight. Two different approaches implemented for navigation purpose such as global and local path planner and social proximity layer technique used for collision avoidance. Four different topologies has been implemented in order to compare behavior of UAV swarm possibilities such as individual decision making, leader follower, predecessor and two nearest predecessor topologies. All of these topologies' performance has advantages and disadvantages according to results extracted from log files by plotting graphs. UAV swarm flight performance has been improved by giving different input parameters such as velocity, acceleration, cost map and searching radius. Different input parameters contributed to improve flight performance in terms of execution time, blocking time etc. Structure: introduction, applications of UAV swarm, classification of UAVs, sensors, UAV swarm communication architecture, path planning, collision avoidance, obstacle avoidance, conclusions and references. Thesis consist of: 78 p. 53 figures, 8 tables and 32 bibliographical entries. |
