Ground Autonomy Evolution - Autonomous and Autonomic Systems - page 85

2. When the Modeler receives the voltage from the Perceptor, it converts the
voltage to a discrete value and updates this value in the model. In this
case, the updated voltage value puts it below the acceptable threshold
and changes the model's voltage state to “low.” This change in state
value causes a state change event and the Modeler now publishes the new
state value to all components that have subscribed to changes in this
state variable. Since the Reasoner has subscribed to changes in this state
variable, the low voltage value is sent to the Reasoner.
3. In the Reasoner, the low voltage value fires a rule in the expert sys-
tem. This rule calls a method that sends the Planner/Scheduler a goal
to achieve a battery voltage level that corresponds to fully charged.
4. When the Planner/Scheduler receives the goal from the Reasoner, it
queries the Modeler for the current state of the satellite and a set of
actions that can be performed (this set may change based on the health
of the satellite).
5. After receiving the current state of the satellite and the set of available
actions from the Modeler, the Planner/Scheduler formulates a list of ac-
tions that need to take place to charge the battery. It then sends the plan
back to the Reasoner for validation.
6. The Reasoner examines the set of actions received from the Planner/
Scheduler and decides that it is reasonable. The plans are then sent to the
Agenda.
7. The Agenda then puts the action steps from the plan into a queue for the
Executive.
8. As the Executive is ready to execute a new step, the agenda passes the
plan steps one at a time to the Executive for execution.
9. The Executive executes each action until the plan is finished. At this time,
the Executive notifies the Agenda that it has finished executing the plan.
10. The Agenda marks the plan as finished and notifies the Reasoner (or
whomever sent the plan) that the plan finished successfully.
11. After the plan is executed, the voltage starts to rise and will trigger a state
change in the Modeler when the voltage goes back into the fully charged
state. At this time, the Reasoner is again notified that a change in a state
variable has occurred.
12. The Reasoner then notes that the voltage has been restored to the fully
charged state and marks the goal as accomplished.
4.3.4 ACT Operational Scenario
The operational scenario that was developed to evaluate ACT was loosely
based on certain nanosatellite constellation ideas. Figure 4.7 graphically illus-
trates this scenario. It was based on the idea of a ground-based community
of proxy agents (each representing a spacecraft in the nanosatellite constel-
lation) that provided autonomous operations of the constellation. Another