Environmental Engineering Reference
In-Depth Information
However, a Remote Agent implementation offers much promise for producing
significant reductions in FSW development and testing costs.
Since each agent can be built as a standalone module (consistent with the
objects associated with object-oriented design), the development schedule for
an agent can be synchronized with the availability of the information needed
to define its requirements. As long as their interfaces with the executive agent
and other agents with which they need to interact with are well-defined, those
agents (for example) associated with hardware components that will only be
procured toward the end of the lifecycle can be developed later than those
agents whose functionality is well-understood from the start.
As agents are developed, they can easily be added to the system, and if
a problem develops with an agent inflight, it can be dropped oine without
H&S impact to platform or payload. As this approach to developing FSW ma-
tures, it will be possible to build up a library of agents from previous missions,
which can be reused economically on future missions once protocols and stan-
dards for agent communication have been established, eventually stimulating
the creation of generalized COTS products, which will greatly facilitate the
reduction of FSW development costs.
Significant reductions in FSW testing costs can also be expected. Since
each applications agent is decoupled from direct communication with the FSW
backbone and since their communication with the backbone is bandwidth lim-
ited by the executive agent, a modification to an applications agent should
not normally require full-scale system-level regression testing. The modified
agent could instead just be tested at a module level and then added back into
the flight system. As an agent can drop oine without impacting the FSW
backbone (and its corresponding H&S functionality), less stringent (and less
costly) testing standards may be applied to the applications agents than to
the backbone and than as currently applied to all conventional FSW. Finally,
the similarity of this software architecture to typical ground system architec-
tures should enable (in some cases) initial agent software development in the
ground system, with the associated cheaper software development and testing
methodologies, with later migration to the flight system following operational
checkout (see Chap.
9
for a detailed example of this, called progressive auton-
omy).
6.6 Mission Types for Remote Agents
In this section, potential advantages of Remote Agents (over a conventional
FSW design) are evaluated at a high level relative to a set of characteristic
mission types, specifically:
1.
LEO celestial pointers
2.
LEO earth pointers
3.
Geosynchronous-earth-orbit (GEO) celestial pointers
