Environmental Engineering Reference
In-Depth Information
2.1.1 Planning and Scheduling
Especially for low earth orbit (LEO) missions, the ground system P&S func-
tion traditionally has been responsible for generating a detailed desired opti-
mized timeline of spacecraft activities, sometimes (as in the case of Hubble
Space Telescope (HST)) based on rather complex predictive modeling of the
spacecraft's environment and anticipated behavior. The ground system then
recasts (and augments) the timeline information in an appropriate manner
for processing on the spacecraft. The timeline is then executed onboard via a
(largely) time-driven processor. Often along with the nominal, expected time-
line, the ground interleaves it with a large array of alternate branches, to be
executed in place of the nominal timeline when certain special conditions or
anomalies are encountered. The resulting product is a highly coupled, time-
dependent mass of data, which, in the past, occupied a substantial fraction of
available onboard storage.
Ironically, the ground system's creation of the timeline data block is it-
self a process almost as highly time-dependent as the execution of the actual
timeline onboard. HST provides a particularly complex example. Long-term
scheduling (look-ahead intervals of several months to a year) was used to
block-out accepted proposal targets within allowed geometric boundary con-
ditions. The geometry factors are typically dominated by Sun-angle consid-
erations, with additional contributions from issues such as moon avoidance,
maximizing target orbital visibility, obtaining significant orbital dark time or
low sky brightness, and meeting linkages between observations specified by
the astronomer.
On the intermediate term (a few weeks to a couple of months), LEO space-
craft targets were ordered and scheduled relative to orbital events such as
South Atlantic Anomaly (SAA) entrance/exit and earth occultations, and the
duration of their associated observations was estimated based on required ex-
posure time computations. Concurrently, support functions requiring schedul-
ing, like communications, were interleaved with the science-target scheduling.
Lastly, on the short term (a day to a week), final detailed scheduling (both
of science targets and support functions) to a precision of seconds was per-
formed using the most accurate available model data (for example, the most
recent predicted spacecraft ephemeris), with the possibility for including new
targets (often referred to as targets of opportunity (TOOs)) not previously
considered.
At times, the software needed to support the intermediate and short-term
scheduling process performed by the ground system has been massive, com-
plex, and potentially very brittle. Further, multiple iterations of the process,
frequently involving a lot of manual intervention (at considerable expense),
were often required to produce an error-free schedule. Although considerable
progress has been made in streamlining this process and reducing its associ-
ated costs, the mathematical modeling remains fairly sophisticated and some
amount of operational ineciency is inevitable due to the necessity of relying
on approximations during look-ahead modeling.
