Autonomy in Spacecraft Constellations - Autonomous and Autonomic Systems - page 183

As in other mission types discussed earlier, the motivations for improved
autonomy in constellations arise from (among other things) resource con-
straints pertaining to onboard processor speeds, memory, spacecraft power,
etc. Even though onboard computing power will increase in the coming years,
the resource constraints associated with space-based processes will continue
to be a major factor that needs to be considered when dealing with, for exam-
ple, agent-based spacecraft autonomy. The realization of “economical intelli-
gence,” i.e., constrained computational intelligence that can reside within a
process under severe resource constraints (time, power, space, etc.), is a major
goal for future missions such as nano-sat constellations , where resources are
even more constrained due to their small size.
Like other missions, satellite constellation missions can have a wide range
of characteristics. Future missions may vary in their data rate and total data
volume. Orbits may range from low earth orbits to very elliptical orbits with
multiday periods. Air-to-ground protocols may vary, and the satellites them-
selves may be low-cost with low autonomy or may be sophisticated with a
high level of onboard self-management. Traditional ground-support systems
designed for single satellite support may not eciently scale up to handle
large constellations. The interested reader may refer to [ 15 , 64 , 143 , 166 , 190 ]
for additional information on the challenges of spacecraft constellations.
Table 9.1 summarizes current and future types of constellation, their ap-
plications, some of the critical distinctions between the applicable mission
models, and various relevant issues. To begin to address the implicit chal-
lenges, new approaches to autonomy need to be developed for constellations.
As discussed in Chap. 4 relative to the agent concept testbed (ACT) proto-
type, a possible two-step approach for achieving constellation autonomy is as
follows:
1. Develop a community of surrogate ground-based agents representing the
satellites in the constellation. This will enable the mission to establish, in
a prototype environment, the centralized and distributed agent behaviors
that eventually will be used in space.
2. Migrate the surrogate agents to the space-based satellites on a gradual or
as-needed basis. This step is referred to as progressive autonomy .
This chapter will focus on Step 1, and at the end, present some ideas relat-
ing to Step 2. First we present a brief overview of constellations, reasons for
using constellations, and the associated challenges in developing them. These
challenges will motivate the agent-based technology discussion in relation to
the goal of achieving autonomy in constellations (Fig. 9.1 ) .
9.2 Constellations Overview
Constellations have the potential to provide the data that are needed to
yield greater scientific insight and understanding into the cause-and-effect
processes that occur in a region. As discussed in Chap. 6 , constellations can