Abstraction morphisms for world modelling in high autonomy systems

A model-based autonomous system uses a multiplicity of models at various control layers to support the predefined system objectives. These models differ in level of abstraction and in formalism. It is noted that concepts and tools are needed to organize the models into a coherent whole. The authors have developed abstraction mechanisms for mapping models of laboratory instruments into operational abstractions, and for mapping a task plan hierarchy into an isomorphic tree of model abstractions. In the present work, they discuss further efforts to develop abstraction mechanisms for systematic derivation of related models through the use of system morphisms. They describe an abstraction mechanism for abstracting spatial relations between the model of the external world and the internal world model used in an autonomous agent. They also show how such a mechanism supports world modeling coherence. Then they illustrate the integration of the world modeling approach and abstraction mechanisms in an autonomous multi-agent laboratory application