An Indoor Knowledge Graph Framework for Efficient Pedestrian Localization

A large amount of information including indoor spatial structure information and smartphone sensor information can be utilized for indoor localization. However, it remains an open challenge to efficiently organize and effectively integrate this data to achieve accurate indoor localization. Knowledge Graphs have powerful intuitive data representation capabilities, semantic and relational processing capabilities, and efficient data storage and feedback mechanisms. Introducing Knowledge Graphs to indoor localization can effectively improve positioning accuracy and subsequently, enrich indoor location-based services. This article presents a Knowledge Graph framework that integrates the basic structure of the indoor environment and various types of smartphone sensing data for indoor localization. This framework consists of two sections: indoor space ontology and mobile sensing data. The indoor space ontology expresses the indoor spatial structure and relationship data, while the sensor sensing data includes a large amount of sensor information generated by pedestrians during indoor activities. Experimental results confirm that the proposed Knowledge Graph framework can achieve efficient indoor localization with good scalability and flexibility under various indoor circumstances.