From consistency to flexibility: shifting the structure, Towards a new generation of geographical information systems

GILLES-ANTOINE NYS | 2023

Summary 

Smart Cities, Enhanced 3D city models, Digital Twins, Small scale city models, Urban Shadows, all combinations of these words, etc. are terms we hear more and more. No matter what they are called, they can be summarised as enhanced 3D city models. Behind these names are hidden concepts, algorithms and systems that focus on opening information and providing smarter tools. Implemented mechanisms should help people to reach an improved level of support in decision-making (crosschecking of information, reasoning, machine learning, etc.). The urban built environment, because of its social, political and economic components, has so many specificities, that it is necessary to propose dedicated management tools in addition to a purely geometric management of the data. The above-mentioned terms translate the increasing need to provide new solutions for an integrated data management in a digital replica of a city. Three pillars are mandatory in order to achieve such an integrated basis: (1) usability, (2) interoperability and (3) maintenance. Things should be understandable in order to open the use by the greatest number and thus provide the full potential of shared information. Some elements could still be considered as advanced and provide additional information and complexity but the base should be kept simple. The second imposes elements to communicate in a standardized way: storages and exchanges should follow an unified schema in order to ease communication between them. Thus, information conveys data understandable by all, and everyone knows where to find the desired information in the model. Constituting stacks should be documented so that the standard can evolve or be modified but still operates in a manner that allows other application to communicate. External connections should also be made in a manner that encourages cross-compatibility. Finally, the third focuses on providing tools that handle data during all its lifecycle and ease its management. From its creation passing by modifications to its destruction, every single step should be handled in order to alleviate problems due to complex processing chains.

This thesis proposes to renew the model storage from a rigid and consistent tabular structure, which are relational databases, to a flattened solution that should improve flexibility (i.e., a document-oriented NoSQL solution). This ambitious goal is structured in the form of a "three-paper" thesis with considerations that place them in the main objective. Chapters are thus divided in three corresponding researches that concern respectively: the data/clients, the database tier and the server tier. The first contribution proposes a simple generation process for the creation of CityJSON buildings models. It develops the benefit of providing a CityJSON model directly and its new supported features (metadata, etc.). The second chapter discusses the creation of a document-oriented schema that improves the city objects and features access thanks to the flattening of the model. A web architecture is proposed enjoying its flexibility. The last chapter focuses on providing a consistent way of writing and reading information using the above-mentioned schema. 3D city models are afterwards seen as a unique integration basis for information coming from heterogeneous sources. Interactions with this single data storage are locked through the provision of a middleware, a software that lies between applications and rules their exchanges.

All of these contributions are packed and staged in a coherent whole that answers a crucial question: Could a flexible web architecture gather the consistency and interoperability aspect of the CityGML data model with better flexibility to cope with a high demand for sharing by applications from different domains? Finally, the conclusion gives an overview on the results of the thesis. Since this thesis should be seen as a part of whole that goes toward a new generation of geographic information systems, research extensions are proposed in order to make steps towards the accessibility of a digital replica of cities. These propositions are made based on current knowledge and technologies; they are part of the present time and cannot assume what will happen next.