Open-Source Flight Management and Its Application to the Flying Fish UAS

A student/faculty team in the University of Michigan Aerospace
Engineering department has developed
an open-source flight management system (FMS) to support autonomous
and semi-autonomous unmanned
air system (UAS) flight. I will overview the fundamental design and
current implementation,
as well as efforts to extend decision-making capabilities beyond the
guidance, navigation,
and control systems required to execute waypoint-based flight
plans. Emphasis will be placed
on the potential to use AI technologies for long-endurance,
highly-complex mission scenarios with
no ground-based supervision.

The Flying Fish unmanned aerial system, the first beneficiary of the
UM flight management software,
is designed to perform persistent surveillance on the open ocean by gathering
data as a drifting surface buoy and periodically repositioning via
self-initiated flight.
Flying Fish maintains a preset watch area, defined by center
coordinates and a radius,
through autonomous flight/drift cycles. I will describe the Flying
Fish mission and multi-mode guidance
and control system that enable fully-autonomous flight operations. I
will present results from a
series of flight tests conducted over the past two years and will
describe ongoing efforts to
extend software "intelligence" to the extent necessary to support
fully-autonomous overnight deployment
planned for July 2009.