Mission Management

Requirements Summary

The Mission Manager is the key to achieving successful integration of the "real" subsystems of the PA.  It was responsible for:

• ensuring communication between the PA subsystems,
• maintaining shared models of the physical environment (the aircraft and the external situation), and
• maintaining a shared model of the pilot's intent - both the static structure, and the dynamic instances of the nodes in the PGG.  

LASC Marietta was the lead agency for developing both the Phase 1 and Phase 2 implementations of the Mission Manager.

Phase 1 Implementation

The Mission Manager was implemented in Phase 1 as an OPS-5 rule system running on a Symbolics Lisp machine. All subsystem communications were sent to the MM which then reformatted and distributed the messages to the other subsystems according to the rules for each type of message. This was extremely useful in integrating subsystems built in remote locations across the country.  The first system integration was performed at Lockheed-Georgia using functional dummies of the real subsystems, making sure that the message rules were able to deliver the messages to the right places.  This complete PA system was then delivered to each of the subcontractors as a baseline.  They were in theory able to replace the dummy of their system with the real subsystem that should supply and consume messages in the same form and sequence as the dummy system.

Key to this approach was the need for an early decision on the scenario events to be implemented, the knowledge to respond to scenario events and the message traffic that would result.

As the systems became more complex and the required discussion became more intense, we migrated to the normal system integration plan in which subcontractors brought their systems to the integration lab and did the integrati0on face-to-face.  However, the need to specify in detail the messages to be produced and consumed, and the scenario to focus the integration was key to every integration in Phases 1 and 2.  There were 12 integration cycles in Phase 1 and 4 in Phase 2.  Each cycle followed this format:

• planning meeting to define the scenario
• written requirements for the knowledge
• parallel development of the systems and the simulation tools to exercise the systems
• integration and test.

Phase 2 Implementation

In Phase 2, the MM was recoded in C++.  Its responsibility for managing communications was eliminated, and replaced by direct subsystem-to-subsystem socket connections.  It also no longer maintained the PGG class structure - that was buried into the knowledge bases of the subsystems.  It did, however, continue to maintain the dynamic PGG structures and the world models of the aircraft and external environment.

last updated 1/26/2003 by David Smith