Environmental Engineering Reference
In-Depth Information
6.5.1 Eciency Improvement
Traditionally, onboard schedule execution has been mostly absolute time-
driven. This design choice, while enabling an extremely simple onboard
element in the overall schedule execution process, has necessitated the de-
velopment of a very expensive ground planning and scheduling element run
by a large, expensive operations staff. The resulting product, due to the high
reliance on the ground system's approximate look-ahead modeling with its
associated worst-case time pads, has also been somewhat inecient and quite
limited in its capacity to respond to off-nominal or unexpected events. The
following discusses cost savings enabled by an onboard scheduling capability.
Event-Driven Scheduling
Autonomous, event-driven management of onboard activity transitions by Re-
mote Agents has the potential for considerable cost savings. There always has
been considerable use of event-driven transition control of onboard processes.
For example, onboard logic can control mode transitions such as slewing to
inertial hold when body rates and pointing errors have dropped below thresh-
old levels, or the transition from inertial hold to safemode when anomalies
are detected requiring immediate, extreme responses. Onboard logic, either
existing within FSW or hardwired into the flight hardware itself, has also
controlled transitions in sensors/SIs from search activities to tracking activ-
ities when the target objects' signal profile warranted it. Still, the use of an
event-driven mechanism for managing transitions between scheduled science
observations has often been precluded by the complexity of the spacecraft's
orbital environment, performance demands imposed by the mission, and the
computational limitations of the flight data system.
An example of an event-driven short-term scheduling system is the Adap-
tive Scheduler originally proposed for use on the James Wells Space Telescope
(JWST). In Adaptive Scheduling's simplest form (from an onboard perspec-
tive), the ground system would still be responsible for generating an ordered
list of desired science targets, but no absolute-time tags would be attached.
The FSW then would observe the targets in the specified order, but nomi-
nally would trigger the execution of the next observation on the list in response
to a FSW event signaling the successful completion of the previous observa-
tion. This avoids the waste of potential observing time engendered by the old
paradigm's use of worst-case time pads to space out absolute-timed commands
that might otherwise “collide” with each other.
The Adaptive Scheduler would also have the capability to insert into the
timeline engineering events, as required, when informed by other FSW subsys-
tems that an action needs to be taken. For example, when angular momentum
has built up to the point that a momentum “dump” must be performed, the
ACS would accordingly inform the Adaptive Scheduler, which would then find
