Exploring a pure landmark-based approach for indoor localisation

Exploring a pure landmark-based approach for indoor localisation

Willems, Oscar (TU Delft Architecture and the Built Environment)

Zlatanova, Sisi (mentor)
Verbree, Edward (mentor)

Delft University of Technology

Geomatics

2017-11-03

Humans interact more and more with their environment through technology, recent decades have seen a huge increase in the need for and availability of Location-Based Services (LBS). Recently landmarks have gotten a renewed interest in the field of LBS , although already a quite old phenomenon. In both the outdoor and indoor environment they are being used to enrich existing services such as navigation, but not used as the basis for a technique or service.

The indoor environment relies heavily on building specific and less-scalable sensor-based localisation techniques (such as Wi-Fi and Bluetooth), alternatives to sensor-based are becoming a necessity and would be a welcome addition. The exploration and development of landmark-based approaches for indoor localisation is something that can extend the field of geomatics and \ac{lbs}.
This research investigates if a pure landmark-based approach works for indoor localisation and which characteristics of landmarks can be exploited. This is achieved by developing a conceptual framework that explores how a landmark-based indoor localisation would work from an artificial point of view. A Minimal Viable Product (MVP) is implemented to evaluate if a landmark-based approach works and what needs to be improved or considered in future studies

Starting from an artificial test case, the MVP to achieve indoor localisation is implementing and evaluated using a manually digitised real-world and more complex test case. The fundamental principle of landmark-based localisation is that through the observation of landmarks within the (indoor) environment a user’s location is obtained because the visibility and location of landmarks are known. The workflow to go from an observation to a location is by 1) calculating the visibility/isovist area of each landmark, 2) interpret the observations into a combination of landmarks, 3) intersect the visibility of all landmarks in the observation, 4) refine the location based on relative landmark constellations, and 5) follow-up with questions on potentially visible landmarks to improve location further

One of the key giveaways of this research is that approach for indoor localisation a landmark-based is feasible, principles and techniques exist (or are being developed), it is only a matter of setting them up in the right order and format them to work, and connect input with the researched process and use them for LBS driven applications.
Future work on the subject of landmark-based localisation and LBS is connecting with existing spatial standards, extend the principles into the 3^rd dimension, and integrate more aspects of landmark salience.

Landmarks
indoor environment
location-based services
localisation
SQL
PostGIS

http://resolver.tudelft.nl/uuid:6332755b-72ea-4a74-a781-5cb3f951de67

Student theses

master thesis

© 2017 Oscar Willems