Environmental Engineering Reference
In-Depth Information
spacecraft revisits are not extremely costly or inecient, maintaining both SI
data processing and TOO scheduling responsibility within the ground system
is probably a better trade than migrating the capabilities to the spacecraft.
However, if the TOO has a very short time duration (relative to turnaround
times for ground SI data processing and scheduling), if target revisits are
costly, or if the spacecraft is not in regular contact with the ground, advantage
can be gained by installing onboard at least limited functionality in these
areas.
For example, after execution of high level survey measurements of a gen-
eral region of the celestial sphere, an SI Remote Agent could process the data
collected and identify point targets appropriate for the follow-up work. That
agent could then inform the planning-and-scheduling agent of the presence of
interesting targets in the immediate neighborhood, which could then sched-
ule and execute those targets immediately (via SI commanding), avoiding
the operational overhead of a re-visit following processing of the survey data
on the ground. So by coupling the efforts of planning-and-scheduling and SI
data-processing Remote Agents onboard, overall system responsiveness can be
greatly improved and measurements of some classes of science targets can be
performed that would otherwise not be observed. This capability has already
been utilized on the Swift mission.
Although this example primarily illustrates the cooperative behavior of the
planning-and-scheduling agent, the SI commanding-and-configuration agent,
and the SI data-processing agent, additional participation by other agents
(including attitude/orbit determination, attitude/orbit maneuvering, and SI
data storage) could also be required.
6.4.4 Goal-Driven Target Scheduling
Traditionally, the ground system has uplinked to the spacecraft a fixed, time-
sequenced target list that the FSW has executed in an absolute time-driven
fashion precisely as specified by the ground. This is probably the optimal
scheduling approach for missions in which the science targets are easily de-
fined in advance and the spacecraft is in regular contact with the ground.
However, for missions more isolated from ground control or where the sci-
ence targets can only be specified by general characteristics and have limited
contact duration opportunities (e.g., asteroid flybys), a goal-oriented target-
specification technique may be essential.
For the case of an asteroid flyby, the SI data-processing Remote Agent
could not only perform straightforward data reduction of the measurements,
it could also employ sophisticated pattern-match techniques (possibly via case-
based reasoning) to identify regions for closer investigation. In this respect,
the previously discussed TOO scheduling is just a subset of goal-driven target
scheduling. However, goal-driven scheduling might also include the onboard
specification of targets without any a priori ground input whatsoever. For
example, suppose a percentage of a spacecraft's mission consists of performing
Search WWH ::




Custom Search