Environmental Engineering Reference
In-Depth Information
The left side of the middle section represents the agents that are ready to
be sent up to a spacecraft, or are operating in a proxy mode. Those agents
that are waiting to be sent to a spacecraft may be ones waiting to be uploaded
for emergency-resolution purposes (e.g., for anomaly situations), or may carry
functionality updates for other agents and are waiting to be uploaded when
needed or when resources become available. The agents that are operating in
a proxy mode are operating as if they were on the spacecraft, but due (for
example) to resource restrictions are temporarily or permanently operating
within the ground control system. The proxy agents communicate with other
agents and components on the spacecraft as if they were running onboard
(subject, as always, to communications constraints).
In a situation requiring an agent to be uploaded, a managing agent in
mission control would make a request to the agent-development area (or the
repository of validated agents) for an agent with the needed capability, and
would (subject to communications constraints) notify the original agent that
requested the capability as to the availability of the requested agent. The
original requesting agent would then factor this information into its planning,
which would be particularly important if the situation were time-critical and
alternate actions needed to be planned if the new agent could not be put into
service in the constellation within a needed timeframe (e.g., as a result of
communications vagaries).
9.6.3 Space-Based Autonomy
The upper part of Fig.
9.5
depicts other communities that are purely opera-
tional on a spacecraft or other robotic system, the members of which would
be mature agents that would have been approved through an appropriate
process. These agent communities may be based around spacecraft subsys-
tem (e.g., instrument, recorder, spacecraft state and management, etc.) or
represent a functionality (e.g., anomaly detection, health and safety, science
opportunity, etc.).
An agent would be able to migrate from its initial community to other
nodes in “agent space” (for lack of a better name). These communities may
be logically or physically distinct from the agent's initial community. A single
agent may either migrate to a new community when it is no longer needed,
or clone itself when needed in multiple communities simultaneously.
The idea of realizing constellation autonomy first through ground-based
communities of spacecraft surrogate agents and then migrating the agent com-
munity to the actual spacecraft is a flexible, dynamic approach to providing
ongoing updates to spacecraft functionality. The progressive autonomy that
could be realized through this approach would enable mission control to up-
load only those agents in the community that have been thoroughly verified
and in which there is the appropriate degree of trust.
