A Novel Approach To The Rendezvous Problem For Aerial Refueling Of Multi-Uav Systems

Advancements in unmanned aerial vehicle (UAV) technology have led to theirwidespread adoption in both industry and academia. In recent years, the increasingfocus on multi-agent systems has brought UAVs further into the spotlight. MultipleUAVs can perform tasks that are beyond the capabilities of a single vehicle orexecute shared missions with enhanced efficiency and effectiveness. Aerial refueling,a strategic and long-established procedure in both civil and military aviation,significantly extends the range and operational capabilities of aerial vehicles. Recentresearch has explored automating this process for both manned and unmannedaircraft; however, current efforts primarily address single-vehicle scenarios. Giventhe growing use of UAV swarms in applications such as search and rescue, terrainmap ping, and disaster relief, autonomous aerial refueling for multiple UAVsrepresents a logical and necessary next step. This thesis addresses the rendezvousproblem for multiple UAVs approaching a tanker aircraft. Reference trajectories aregenerated using virtual spring-damper and charged particle models. A distributedcontrol architecture is designed, employing waypoint-following controllers for individualvehicles. In 2D proof-of-concept scenarios, non-holonomic UGVs are utilized,and their motion control is achieved via cascaded controllers which utilize virtualinputs. Additionally, a novel collision avoidance algorithm and a two-phase dockingprocedure are proposed. To validate the proposed framework, a series of coordinatedtask scenarios involving groups of UGVs and UAVs are simulated in both 2D and3D environments. The results confirm the effectiveness of the coordination strategyand formation control, demonstrating robust and satisfactory performance across avariety of scenarios.