Define proper objective status logic (state machine?)

[chcorbato]

mc_mros_reasoner/mros2_reasoner/mros2_reasoner/ros_reasoner.py

Line 381 in 9698265

if obj_in_error.o_status in ["UPDATABLE"]:

This needs to be fixed simultaneously with logic in the Wrapper (monitors objectives) and in the MROS Observer (which apparently in ROS2 monitors components/nodes)

[estherag]

@chcorbato I've been thinking how to solve this but it is not that easy. The problem is that the Observer only handles components and the Wrapper only handles objectives. When a component is RECOVERED we have two options:

• Extend the observer to also handle objectives:

  • Send a objective UPDATABLE status and handle it as if the Objective were in INTERNAL ERROR -> need to select the best, available and realizable FD to ground

• Keep the structure as it is:

  • The observer sends a component RECOVERED status to the ontology
  • We apply a rule to set the objective to UPDATABLE status
  • For the next loop, the component shall be in NONE status (not updatable, not in error) so in Python we remove the status of the component

In both cases, once a FD is grounded, the objective status is reset to NONE in the ontology