Login / Signup

Model-driven engineering city spaces via bidirectional model transformations.

Ennio ViscontiChristos TsigkanosZhenjiang HuCarlo Ghezzi
Published in: Software and systems modeling (2021)
Engineering cyber-physical systems inhabiting contemporary urban spatial environments demands software engineering facilities to support design and operation. Tools and approaches in civil engineering and architectural informatics produce artifacts that are geometrical or geographical representations describing physical spaces. The models we consider conform to the CityGML standard; although relying on international standards and accessible in machine-readable formats, such physical space descriptions often lack semantic information that can be used to support analyses. In our context, analysis as commonly understood in software engineering refers to reasoning on properties of an abstracted model-in this case a city design. We support model-based development, firstly by providing a way to derive analyzable models from CityGML descriptions, and secondly, we ensure that changes performed are propagated correctly. Essentially, a digital twin of a city is kept synchronized, in both directions, with the information from the actual city. Specifically, our formal programming technique and accompanying technical framework assure that relevant information added, or changes applied to the domain (resp. analyzable) model are reflected back in the analyzable (resp. domain) model automatically and coherently. The technique developed is rooted in the theory of bidirectional transformations, which guarantees that synchronization between models is consistent and well behaved. Produced models can bootstrap graph-theoretic, spatial or dynamic analyses. We demonstrate that bidirectional transformations can be achieved in practice on real city models.
Keyphrases
  • physical activity
  • mental health
  • healthcare
  • primary care
  • magnetic resonance imaging
  • magnetic resonance
  • health information
  • computed tomography
  • data analysis
  • convolutional neural network
  • image quality