Model-Driven Design of City Spaces via Bidirectional Transformations

Technological advances enable new kinds of smart environments exhibiting complex behaviors; smart cities are a notable example. Smart functionalities heavily depend on space and need to be aware of entities typically found in the spatial domain, e.g. roads, intersections or buildings in a smart city. We advocate a model-based development, where the model of physical space, coming from the architecture and civil engineering disciplines, is transformed into an analyzable model upon which smart functionalities can be embedded. Such models can then be formally analyzed to assess a composite system design. We focus on how a model of physical space specified in the CityGML standard language can be transformed into a model amenable to analysis and how the two models can be automatically kept in sync after possible changes. This approach is essential to guarantee safe model-driven development of composite systems inhabiting physical spaces. We showcase transformations of real CityGML models in the context of scenarios concerning both design time and runtime analysis of space-dependent systems.