BIM-GIS-IoT-based semantic pedestrian network modelling and applications in high-density cities

Abstract

In high-density cities, pedestrians frequently traverse many publicly accessible indoor spaces, such as metro stations and footbridges, which connect with outdoor sidewalks, forming indoor-outdoor combined pedestrian networks. Simultaneously, the Internet of Things (IoT) sensors capture substantial data for monitoring dynamic changes in the walkable built environment. Although they are physically connected, their connections in the digital world are unestablished, leading to difficulties in obtaining integrated information for intelligent pedestrian services, such as indoor-outdoor combined route planning, integrated information queries and real-time risk identification. Currently, outdoor sidewalks and indoor spaces are modelled separately using Geographical Information Systems (GIS) and Building Information Modelling (BIM) technologies, while sensor data is stored in diverse IoT databases. Multi-source data employed by different technologies with various data schemas and storage methods hinder information integration. Therefore, establishing indoor-outdoor combined pedestrian networks in the digital world by integrating BIM, GIS, and IoT data is crucial. This research aims to model the BIM-GIS-IoT-based semantic pedestrian network (SPN) and design applications of the SPN using semantic web technologies. Firstly, a BIM-based indoor SPN is modelled by analysing the BIM data schema, extending the indoor SPN-oriented ontology, and developing a BIM data-to-linked data converter. The resulting indoor SPN is a linked data-based graph, serving as a knowledge base for indoor pedestrian services. Secondly, a BIM-GIS-based indoor-outdoor combined SPN is modelled by establishing interlinks between cross-domain ontologies and instances from BIM and GIS datasets using semantic queries and logical inferences. The resulting indoor-outdoor combined SPN serves as a knowledge base for indoor-outdoor integrated services. Finally, a BIM-GIS-IoT-based dynamic SPN is modelled, which mainly involves converting IoT sensor data into linked data and interlinking instances from BIM/GIS/IoT datasets. The resulting dynamic SPN integrates static spatial data from BIM and GIS datasets with dynamic information from IoT databases, thus serving as a knowledge base for real-time pedestrian services. To validate the resulting SPN, various application scenarios are designed and demonstrated at each stage. Case studies are conducted using BIM models for an underground metro station and a university campus building, a GIS model for campus sidewalks, and IoT time-series data. The information queries are tested in each stage to demonstrate the generated SPN compliance with the predefined ontologies and linked data syntax. Subsequently, indoor-outdoor combined route planning is devised, leveraging a navigation-oriented subgraph derived from the SPN. Moreover, automatic risk identification and updates are achieved by defining semantic rules for indoor fire, outdoor flooding, and crowd congestion risks. The results of case studies revealed that the SPN promotes integrated and informed decision-making, integrates cross-domain information, facilitates the development of relevant knowledge bases using BIM-GIS-IoT data, and enhances intelligent pedestrian services in high-density cities. The significance of this study lies in the proposed modular approach for integrating multi-source data, serving as a foundation for future research in developing city information models and semantic digital twins. Furthermore, the SPN, functioning as a knowledge base harnessing BIM, GIS, and IoT data, holds the potential for broad applicability across various scenarios.