Environmental Engineering Reference
In-Depth Information
one GEO will be integrating and processing data, and a third GEO will be
available for planning and scheduling work, which in this case was assigned to
GEO-4, leaving GEO-1 available to respond to other ground requests as well
as communications from GEO-4.
Once GEO-4 has completed its planning and scheduling work, it commu-
nicates its results to GEO-1. GEO-1 then passes the plan for installing the
new calibrations and procedures to the other GEOs. Each GEO then instructs
those LEOs under its control as to the changes that should be made. The time
at which those changes should be made is also specified to ensure that when
the next mass of data from the whole constellation is integrated, all data in
the mix would have been generated, ideally, using the same calibrations and
procedures. If that ideal condition is not possible, then any data obtained
using old calibrations/procedures will, when contributed to integration, be so
identified in the data/results transmitted to the ground.
B.5 Agent-Based Satellite Constellation Control
Scenario
Consider the scenario when one has many satellites with different viewing ca-
pabilities (IR, visible, or UV) orbiting a planet and one desires a full spectrum
sweep of a certain portion of the planet. Traditionally, science team members
and human controllers would need to identify the satellites with each different
capability that will be making a pass over the section of the planet indicated.
Human controllers would need to form a detailed, possibly quite intricate plan
for the needed observations and to organize a series of requests addressed to
the satellites to perform the sweep, with all relevant details down to the trans-
mission of the data back to earth. This type of activity entails ineciencies
and represents a questionable or wasteful use of manpower.
Through the use of an agent community that hierarchically and intelli-
gently parses instructions, this could be done much more eciently. Ideally,
the human operator needs only to transmit a command similar to “Take a full
spectrum picture of the area bounded by given latitude and longitude data and
transmit the picture back in 2 days.” An executive level satellite could receive
this information, decompose it, and then negotiate with the agent community
(where each satellite in orbit is part of the community) to attempt to sched-
ule a plan. From there, each satellite could respond with information such as
“will be passing over the site in 36 h, I can take the picture in IR” or “will not
be passing over the site for another 96 h, I cannot take the picture.” Certain
constraints may come into play also; for example, UV and visible light sensors
are only useful when it is not night time at the given site. Responses of this
nature may be similar to “will be passing over the site in 4 h, but the site is
currently on the dark side of the planet” or “will be passing over in 4 h when
site is on dark side, but will pass over again in 20 h when it is local noon.”
