Unmanning UAVs – Addressing Challenges in On- Board Planning and Decision Making

Abstract. Planning and decision making, especially the planning of dynamically negotiable collision free paths, is an integral part in the operation of Unmanned Aerial Vehicles (UAVs). Effective path planning ensures that the UAV operates safely, and conforms to the rules and regulations governing flight within the National Airspace System (NAS). To demonstrate an Equivalent Level Of Safety (ELOS) to that of piloted aircraft for certification purposes, UAVs must demonstrate a high level of autonomy without a human in the loop. This research surveys the literature as to how human experts perform planning tasks and forms a framework which promotes shared authority of UAV mission (re)planning and path planning, and can adopt sole authority should the UAV communications link fail or the human operator relinquishes decisions. It has been demonstrated through simulation that the optimization of flight manoeuvre sets using multiple objectives allows for convergence to a solution which better represents civilian mission requirements whilst emulating common flight patterns of trained pilots. These initial findings highlight the challenges involved in replicating the skills of human pilots onboard a UAV. It is revealed that UAV planning and decision making is a multi-disciplinary problem that combines the fields of path planning (search optimization), trajectory generation, and human cognition