Emergency Trajectory Management for Transport Aircraft

Abstract

Over the years gradually flight management systems have been added to airplanes which reduce pilot workload and increase safety. These systems however do not provide an adequate response when in an emergency situation the aircraft loses all thrust. A system that creates glide trajectories to airports that are reachable would enable pilots to focus their attention on stabilizing and regaining control in an in-flight emergency situation. The objective of this study is to develop such an emergency flight system that can generate glide trajectories from the position at which the emergency starts to all reachable runways. By first determining the maximum range of the aircraft, a footprint is defined in which the reachable runways are identified. To reduce risk on board as well as on the ground, the trajectories generated must be balanced between flying over populated areas effectively increasing the risk of loss of life on ground or avoiding populated areas possibly increasing the risk of losing control of the aircraft by having to be airborne for a longer period of time. The trajectories to all reachable runways are ranked based on several airport and runway quality factors and are presented to the pilot to choose from. The trajectories are generated by an Approximate Dynamic Programming (ADP) algorithm which is commonly used in robotics for path planning. In this research it is investigated if it can also be applied in the field of aeronautics for the purpose of an on line emergency trajectory planner. In a test scenarios, trajectories are generated by several ADP variants and compared with a benchmark trajectory created by the A* algorithm. This algorithm generates optimal obstacle avoiding trajectories providing certain conditions are met. The ADP variants that provides the best result is chosen to generate the trajectories for the emergency trajectory planner in a scenario in which the aircraft has to generate trajectories to the reachable airports in the footprint. From the results it could be concluded that approximate dynamic programming can successfully create trajectories to all reachable airports in the footprint. It however proves less suitable to create trajectories on line in emergency situations because the time the algorithm needs to create trajectories is too long to be able to implement it in emergency situations.