DESIGN AND DEVELOPMENT OF A MULTI-SENSOR INDOOR NAVIGATION SYSTEM TO SUPPORT STUDENTS INSIDE CAMPUS BUILDINGS

Global navigation satellite systems (GNSSs) are generally not suitable for indoor navigation due to the lack of good satellite coverage inside buildings. Besides, a robust and accurate positioning system for indoor navigation may involve high costs for local building installations and advanced mobile equipment. Additionally, such systems — especially mobile frontends — are often highly tailored to a particular building and cannot be re-used for other buildings. Hence, a solution is needed that requires neither a costly installation inside buildings nor particular mobile equipment. Moreover, mobile frontends (e.g. mobile applications) should be usable for different buildings. We contribute an indoor navigation system that requires only commodity hardware components. Users can bring their own smartphone equipped with a mobile navigation application. This mobile navigation application provides an import interface for models of different buildings. In turn, the facilities must only provide WiFi access points and/or Quick Response (QR) codes as a minimal localization infrastructure as is often the case. The developed indoor navigation system exploits different sensors (e.g. WiFi, [pseudo]compass, barometer, image recognition) for positioning the users and can detect the direction of (vertical) movement. The mobile navigation application also provides different methods for the localization (e.g. network ID, trilateration, fingerprinting, and weighted average). Moreover, it can calculate routes (distances) between the current and the desired position of a user and track how users move through the building. It supports routing through different floors as well.

The designed solution provides for the flexible management of map data (building plans, floor plans, etc.). Institutions (e.g. universities, public authorities) that meet the mentioned infrastructural requirements can easily create a model of their building by using state-of-the-art building plans supplemented by the location of WiFi access points and/or QR codes. An automatic mechanism extracts a formal graph-like model of the building from the annotated building plans. This is subsequently imported to the mobile navigation application. The developed mobile navigation application works as an interpreter. Thus, the participating institutions need no programming skills to adapt the system for their buildings. Building plans (e.g. scanned documents) can be annotated with open-source standard software.

The designed mobile navigation application is evaluated within a case study. Two campus buildings (more than 500 rooms overall) were modeled and the different localization methods were tested. The evaluation demonstrates the applicability and usefulness of both the modeling approach of buildings and the mobile navigation application itself.