Framework and tool support to design high quality domain models for intelligent systems

Increasing complexity of Cyber-Physical Systems (CPS) is mandating architectures that are autonomous and adaptive to the changing contexts. The area of planning and acting in Artificial Intelligence (AI) provides a well-understood paradigm for designing such intelligent systems. Capturing the domain knowledge formally, both declarative and procedural is crucial to this approach. There have been many attempts to provide assistance for domain designers to specify and develop high quality domain models. However, there has been no conceptual framework to identify all the necessary enablers for this purpose. Also, almost all the existing tools emphasize on visualization tasks and by using simulation. In this paper, we outline a classification framework based on the 3 C's (correctness, completeness and consistency) of specification and identify the design tasks, motivating them through illustrative snippets. We also report a planning and acting tool integrated with simulation and validation capabilities to support these tasks and present its applications that allows experimenting with domain.