Autopilot systems are capable of reliably following flight plans under normal circumstances, but even the most advanced flight-management systems cannot provide robust response to most anomalous events including in-flight failures. This paper describes an emergency flight-management architecture that can be applied to piloted or autonomous aircraft, with focus on the design and implementation of an adaptive flight planner (AFP) that dynamically adjusts its model to compute feasible flight plans in response to events that degrade aircraft performance. A two-step landing-site selection/trajectory generation process defines safe emergency plans in real time for situations that require landing at an alternate airport. A constraint-based search algorithm selects and prioritizes feasible emergency landing sites, then the AFP synthesizes a segmented trajectory to the best site based on postfailure flight dynamics. The AFP architecture is general for any failure situation; however, operational success is guaranteed only with accurate postfailure performance characterization and a trajectory generation strategy that respects degraded flight envelope boundaries. A real-time segmented trajectory planning algorithm and case study results are presented for total loss of thrust failure scenarios. This emergency is surprisingly common and necessitates an immediate approach and landing without a go-around option.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below