Abstract

Unmanned Aerial Vehicles (UAVs) can carry out more complex civilian and military applications with less cost and more flexibility in comparison of manned aircraft. Mid-air collision thus becomes profoundly important considering the safe operation of air transportation systems, when UAVs are increasingly used more with various applications and share the same airspace with manned air vehicles. To ensure safe flights, UAVs have to configure Sense and Avoid (S&A) systems performing necessary maneuvers to avoid collisions. After analyzing the manner of S&A system, avoidance strategies based on a subset of possible collision scenarios are proposed in this thesis. 1) To avoid a face-to-face intruder, a feasible trajectory is generated by differential geometric guidance, where the constraints of UAV dynamics are considered. 2) The Biogeography Based Optimization (BBO) approach
is exploited to generate an optimal trajectory to avoid multiple intruders’ threats in the landing phase. 3) By formulating the collision avoidance problem within a Markov Decision Process (MDP) framework, a desired trajectory is produced to avoid multiple intruders in the 2D plane. 4) MDP optimization method is extended to address the problem of optimal 3D conflict resolution involving multiple aircraft. 5) Considering that the safety of UAVs is directly related to the dynamic constraints, the differential flatness technique is developed to smoothen the optimal trajectory. 6) Energy based controller is designed such that the UAV is capable of following the generated trajectory.