Modeling non-functional properties for human-machine systems

A Human-Machine system is a complex system consisting of many components and services that dynamically compose to achieve a specific goal. The functional and non-functional attributes of the components are considered to make 'who does, what, and when' decisions depending on the operational context. However, non-functional properties are not given sufficient importance compared to that of the functional requirements during the developmental stages. This paper highlights the importance of non-functional properties in human-machine systems and proposes a metamodel for modeling those properties. A case study on assistive lane keeping in automobiles is presented to demonstrate how the non-functional properties can be modeled. This is a part of the intermediate results of a research in progress for modeling decision architectures for autonomous systems.