Environmental Engineering Reference
In-Depth Information
6
Agent-Based Spacecraft Autonomy Design
Concepts
In this chapter, we examine how agent technology might be utilized in flight
software (FSW) to enable increased levels of autonomy in spacecraft missions.
Again, as stated in the Preface, our discussion relates exclusively to uncrewed
assets (robotic spacecraft, instrument platforms on planetary bodies, robotic
rovers, etc.) or assets that must be capable of untended operations (e.g.,
ground stations during “lights-out” operations). The basic operational func-
tionality discussed in Chap.
2
is accounted for and allocated between flight and
ground, and between agent and nonagent FSW. Those technologies required
to enable the FSW design are touched upon briefly, and new autonomous
capabilities supportable by these technologies are described. Next, the advan-
tages of the design from a cost-reduction standpoint are examined, as are the
mission types that might benefit from this design approach.
For each design concept, the appropriate distribution of functionality be-
tween agent and nonagent components and between flight and ground systems
is discussed. For those functions assigned to remote agent processing, enabling
technologies that are required to support the agent implementation are iden-
tified. Further, we determine for which mission types each design concept is
particularly suitable and describe detailed operational scenarios depicting the
remote agent interactions within the context of that design concept.
6.1 High Level Design Features
The general philosophy of this FSW concept is that conventional nonagent
software will encompass all H&S onboard functionality. This FSW segment,
referred to as the “backbone,” will also contain functionality directly support-
ing H&S needs, such as slewing and thruster control, and will be directly ac-
cessible by realtime ground commands in the “normal” manner. The Remote
Agent software will embody mission support functionality, such as planning
and scheduling, as well as science data processing. To achieve its goals, a Re-
mote Agent may access backbone functionality through a managed bus whose
