Developing a Methodology for the Automatic Generation of Geometric Network Model for Indoor Navigation

The demand for indoor navigation services in large-scale indoor spaces is increasing due to its potential applications in efficient wayfinding, emergency evacuation. However, the high cost of indoor spatial data, especially the manual generation of network data, creates inefficiencies and is a major barrier to the commercialization of indoor navigation services. To overcome these challenges, this paper proposes a methodology that automatically generates a GNM (Geometric Network Model) based on NRS (Node-Relation Structure) of IndoorGML optimized for indoor navigation. This paper classified the corridors of indoor spaces into simple polygons and simple polygons with a hole and designed a medial axis extraction and linearization algorithm suitable for each type. In particular, the medial axis extraction using the Voronoi Diagram, polygon subdivision, and medial axis integration process minimized the structural complexity and refined the network data. The proposed methodology enables the automatic generation of indoor space networks from simple input data and is scalable for application to various types of indoor environments, including large-scale spaces. We expect that this methodology will serve not only as a foundational technology for indoor navigation services but also for a wide range of indoor spatial analysis applications.