Environmental Engineering Reference
In-Depth Information
appropriately and with smaller contingency safeguards or margins. Also, a
side benefit to this new paradigm is that wherever expensive ground-system
look-ahead modeling can be replaced by simple onboard realtime response,
cost savings are achieved within the ground system as well.
The example cited in this subsection, the originally proposed JWST Adap-
tive Scheduler, is a rather modest effort at exploiting the potential offered
by an event-driven FSW operation. The Remote Agent FSW formulation de-
scribed in this section would have much more powerful capabilities. More than
just being an event-driven executor of a ground-planned, ground-scheduled
(both long-term and short-term) target list, a Remote Agent design provides
a framework for migrating short-term scheduling responsibility to the onboard
system, greatly increasing the spacecraft's capacity for managing onboard re-
sources more eciently, reacting to anomalies without loss of observing time,
and responding to TOOs while the opportunity presents itself.
Short-Term Scheduling
Also, as discussed for the Adaptive Scheduler's event-driven scheduling,
short-term scheduling by the FSW must be able to accommodate realtime
commanding originating from the ground system. For onboard short-term
scheduling, the complications will be even greater as there is no reason to
expect the ground to be cognizant of ongoing onboard activities when the
ground's command is sent. To deal with such intermittent interrupts of vary-
ing levels of priority, one can envision an onboard “fuzzy logic” reasoning
module that juggles an array of priority levels and time windows attached
to ordered/unordered target lists, target clusters, ground-originated realtime
commands, ground-originated absolute-timed commands with/without win-
dowing, onboard-originated housekeeping/engineering activities, realtime on-
board sensor measurements, and externally-originated realtime data, and then
deduces which activity (or activities) should be executed next (or in parallel).
Further, one could envision an event-driven Remote Agent engaging in a
dialog with more powerful computer resources (and models) on the ground to
acquire deeper look-ahead information prior to making its activity transition
decisions. The agent could even establish dialogs with other spacecraft to
benefit from their realtime measurements. For example, if two earth-pointing
spacecraft were flying in formation, the one arriving later over the target
could interrogate the earlier one regarding cloud cover and other conditions
and make appropriate SI reconfigurations prior to arrival.
6.5.2 Reduced FSW Development Costs
Conventional FSW is highly integrated code, optimized for timing perfor-
mance and computational eciency. As a result, long-term FSW mainte-
nance tends to be very expensive (relatively small code updates may require
very wide-spread regression testing) and FSW reuse has been rather limited.
