Towards Robust Methods for Indoor Localization using Interval Data

Abstract

Indoor localization has gained an increase in interest recently because of the wide range of services it may provide by using data from the Internet of Things. Notwithstanding the large variety of techniques available, indoor localization methods usually show insufficient accuracy and robustness performance because of the noisy nature of the raw data used. In this paper, we investigate ways to work explicitly with range of data, i.e., interval data, instead of point data in the localization algorithms, thus providing a set-theoretic method that needs no probabilistic assumption. We will review state-of-the-art infrastructure-based localization methods that work with interval data. Then, we will show how to extend the existing infrastructure-less localization techniques to allow explicit computation with interval data. The preliminary evaluation of our new method shows that it provides smoother and more consistent localization estimates than stateof-the-art methods.

Handheld Device-Based Indoor Localization with Zero Infrastructure (HDIZI)

“…Fingerprinting [65,66] techniques for indoor localization generate images of objects by sensing electric currents. Then, the image is compared with those in a database (DB).…”

A Safe, Efficient and Integrated Indoor Robotic Fleet for Logistic Applications in Healthcare and Commercial Spaces: The ENDORSE Concept

“…Besides, most of these approaches require initial robot position to be known at least approximately, hence cannot solve the lost robot problem. We will develop a more robust and effective alternative, which is offered by set membership estimation and data fusion techniques that work in the bounded-error framework [5], [6]. These techniques can operate efficiently with sparse, asynchronous and heterogeneous measurements while being robust to the presence of non-consistent measurements, inaccuracy in environment modelling, and drift and inaccuracy in the robot evolution model.…”