Environmental Engineering Reference
In-Depth Information
6.2.6 “Smart” Fault Detection, Diagnosis, Isolation,
and Correction
Some “primitive” fault detection driven by simple rule-based limit checking
must be present in the FSW backbone to control entry into safemode and
other critical mode transitions and/or hardware reconfigurations. However,
that portion of fault detection solely concerned with threats to performance
of science observations (as opposed to the hardware itself) would be assigned
to a Remote Agent. That same agent could also perform fault diagnosis and
isolation and determine the appropriate course of corrective action, which
could be either a predetermined “canned” solution or an original solution in-
dependently created by the agent. To the extent that fault diagnosis, isolation,
and correction exist within the backbone, they should be viewed as canned
correlations between simple limit checks and canned corrective actions devel-
oped and specified by ground personnel, primarily before launch. This agent
may also utilize a wide variety of AI products to generate and interpret its
results, including state modeling, case-based reasoning, and neural nets.
6.2.7 Look-Ahead Modeling
The FSW backbone is a time-driven software element that looks at data in
realtime and responds in realtime or near-realtime: there is no need for any
look-ahead modeling functions within the backbone. On the other hand, many
Remote Agent applications (such as planning and scheduling, data trending
and monitoring, and orbit determination and maintenance) may require the
support of look-ahead models. Examples include ephemerides, solar intensity,
wheel speed, and SAA entrance/exit. The model outputs could be generated
continuously independent of the need of any other agent for their use. These
solutions could then be stored in a data “archive” pending a request by other
agents.
6.2.8 Target Planning and Scheduling
Traditionally, FSW has been time-driven, and full planning and scheduling
responsibility rested with the ground system. The ground would generate an
absolute-timed target list that the FSW would execute precisely as specified.
If conditions were inappropriate for the science observation to execute (for
example, if no guide stars had been found), the spacecraft would still remain
uselessly at that attitude until the pointer in its C&DH FSW moved to the
time of the slew to the next target. However, as planning and scheduling capa-
bilities are migrated from the ground to the spacecraft, more flexible responses
to anomalies are enabled leading to greater overall operational eciency at
reduced costs. It is these new capabilities, which are unnecessary to H&S
maintenance within the FSW backbone, that are the purview of the target
