Modelling Changes, Stakeholders and Their Relations in Semantic 3d City Models

Abstract

Abstract. Urban digital twins have been increasingly adopted by cities worldwide. Digital twins, especially semantic 3D city models as key components, have quickly become a crucial platform for urban monitoring, planning, analyses and visualization. However, as the massive influx of data collected from cities accumulates quickly over time, one major problem arises as how to handle different temporal versions of a virtual city model. Many current city modelling deployments lack the capability for automatic and efficient change detection and often replace older city models completely with newer ones. Another crucial task is then to make sense of the detected changes to provide a deep understanding of the progresses made in the cities. Therefore, this research aims to provide a conceptual framework to better assist change detection and interpretation in virtual city models. Firstly, a detailed hierarchical model of all potential changes in semantic 3D city models is proposed. This includes appearance, semantic, geometric, topological, structural, Level of Detail (LoD), auxiliary and scoped changes. In addition, a conceptual approach to modelling most relevant stakeholders in smart cities is presented. Then, a model - reality graph is used to represent both the different groups of stakeholders and types of changes based on their relative interest and relevance. Finally, the study introduces two mathematical methods to represent the relevance relations between stakeholders and changes, namely the relevance graph and the relevance matrix.

Smart City Digital Twin Framework for Real-Time Multi-Data Integration and Wide Public Distribution

“…Former data and computed representation need to be prepared and distributed in a suitable manner to be accessible in the frontend where the SCDT is presented to decision makers and urban planners, and with limited capabilities to a wide public, too. Implemented interactions have to include the capability to change the point of view (R19), to select entities to access additional information (R21.i, R21.ii, R22, R26) [41], to load different models (R21.iii) [42], and to act on lighting (R20) [43]. To provide more immediate interpretation of visualized data, entities such as PINs (to indicate POI, devices, dehors, etc., on maps), shapes, paths, etc., should be represented with characteristic visualizations (R16) [44].…”

Path-Tracing Semantic Networks to Interpret Changes in Semantic 3d City Models

“…(Nguyen et al, 2017) then proposed a new approach to comparing arbitrarily large CityGML datasets using a graph database. On the other hand, (Nguyen and Kolbe, 2021) explained the importance and advantage of not only identifying but also understanding and evaluating changes with respect to different groups of stakeholders. In general, the major problems of solving change detection and interpretation in semantic 3D city models can be summarized as follows:…”

“…The relevance relations between changes and stakeholders can then be expressed using a table. (Nguyen and Kolbe, 2021) proposed a refined model of changes as well as stakeholders in the context of urban digital twins with semantic 3D city models as one of their key components. The research further provided two mathematical ways to flexibly model the varying relevance relations between different groups of stakeholders and different types of changes, namely using the relevance matrix as well as the relevance graph.…”

“…The context information extracted and aggregated in previous layer L1 as explained in Section 3.2.1 can be forwarded to the Change Type Layer L2 to further classify changes, provide scope and therefore potentially identify local, clustered or global (systematic) changes in the dataset. Conceptual models of changes that are compatible with semantic 3D city models, such as proposed by (Nguyen and Kolbe, 2021), can be used to populate the vertices in this layer. The framework proposed in this research allows flexible structuring of each layer, meaning that entire layers can be reorganized into a corresponding graph representation of a class inheritance hierarchy as long as they satisfy the requirements of the network given in Section 3.1.…”

“…as individual beings (citizens), companies (private sectors) and organizations (non-governmental and governmental organizations), etc. Similarly to previous layers, a conceptual model for actors in layer L4 and stakeholders in layer L5, such as proposed by (Nguyen and Kolbe, 2021), can be employed to populate the network.…”