Approach to uav swarm control and collision-free reconfiguration

Abstract

In this paper, the problem of unmanned aerial vehicle (UAV) swarm control is considered. As the most promising approach to UAV swarm control a composite one was chosen. The proposed conceptual model of the composite method includes operator, base station, and two communication networks: one between agents and another one between base station and agents. UAV trajectory computation is fulfilled on the land-based station using swarm reconfiguration algorithm. Two enhanced 3D reconfiguration algorithms were presented: eased and based on divide and conquer—which differ in used collision avoidance methods. Eased algorithm contains a method of parallel translation of target points, homothety, and target coordinate exchange method. To ensure control of greater swarm size, algorithm based on divide and conquer was developed. Algorithm based on divide and conquer includes parallel translation, delay allocation method, collision graph composition and divide and conquer algorithm. For evaluation algorithm performance tests in virtual environment were performed on swarms of different sizes (from 10 to 100 agents). Greater trajectory calculation time and swarm reconfiguration time are explained by the greater number of used methods. However, the total trajectory and the average number of collisions calculated using algorithm based on divide and conquer have smaller values compared to eased algorithm. Eased algorithm provides UAV movement without collisions with a swarm size of 70 agents, while algorithm based on divide and conquer which has a limit of 90 agents. Therefore, for swarm reconfiguration with a large number of agents, algorithm based on divide and conquer is more preferable.