Figure 1. The ASTRO and NextSat Spacecraft, on Orbit Without the Separation Ring.
It is often the case that new technology for space needs to be taken out of the lab and proven in space. The primary goal of a technology mission
is to prove the capabilities of newly developed space
technology. Most technology missions have the challenge of discovering the limits and capabilities of
new systems, and the Defense Advanced Research
Projects Agency’s (DARPA) Orbital Express (OE) mission was no exception. Orbital Express launched
March 8, 2007, and decommissioned on July 22,
2007. The Orbital Express mission demonstrated on-orbit servicing of spacecraft, including rendezvous,
transfer of battery and CPU modules, and transfer of
propellant, the actual duration of each being approximated but not known to high enough fidelity to
commit to a communications and operations plan.
This introduced challenges because pertinent
information needed for planning was not available
until the various technology experiments were per-
formed, but to perform these experiments we needed
to submit plans to reserve communications
resources, configure execution scripts, and monitor
execution. We note that it is often the case that
bounds of performance are known a priori. Our auto-
mated planning system leveraged this information to
help address the inherent uncertainty in experiment
planning.
The Orbital Express planning cycle consisted of a
long-term plan (planned four weeks in advance) and
a short-term plan (planned the day before opera-
tions). The long-term plan was produced before any
of the experimental/unknown information was
learned. The role of the long-term plan was to ensure
that enough resources are reserved ahead of time.
Planning at this point required accommodation of
the bounds of possible execution. Traditionally, this
level of planning is relatively conservative. For exam-
ple, if we thought it likely that a propellant transfer
would require 20 minutes, but could take up to an
hour, we would plan for the whole hour. We also
couldn’t count on getting all of the communications
passes we asked for, so we would plan for 20 percent
of them to be dropped, and thus ask for more than
we needed.
The short-term plan was produced immediately
before execution, and some of the experimental or
