AI Seminar
Autonomy Augmentation to improve
Add to Google Calendar
Commercial aviation has emerged as a reliable and efficient form of transportation. Although high-profile accidents occasionally occur, air travel is widely regarded as safer than driving an automobile. We have achieved an operational steady state where pilots monitor aircraft systems and coordinate operations through air traffic controllers but leave routine flying to the autopilot. However, current autonomy technologies are somewhat rigid, with limited ability to adapt and accurately make decisions when unanticipated situations arise. This presentation will begin by reviewing the progressive introduction of automation into the aircraft cockpit and into the UAS (Unmanned Aerial System). I will describe ongoing research to further improve safety in the cockpit during situations where aircraft systems fail or the airframe is damaged. As our physics-based models and software algorithms become increasingly complex, automation algorithms are further challenged to provide the pilot with sufficient situational awareness. UAS and heavy transport aircraft are also expected to share the same airspace over the next decade, and the air transportation community is struggling to increase passenger throughput (thus traffic density) at major airports. Military strategists are struggling to define and understand the roles of humans and automation in the cockpit as well as on the ground. As a somewhat controversial alternative to conventional air traffic runway queues, I will overview research in airspace planning for runway-independent aircraft, and discuss the potential for more efficiently (and someday autonomously) managing both terminal area and battlespace traffic with potentially diverse mission goals and performance characteristics.