anticache

Automated Optimal Placement of Anchors in UWB Indoor Positioning Systems (Masterthesis)

27.06.2017

Bernhard Häfner


27.6., 14:00, room 3999

Abstract:

Especially in industrial contexts, an accurate determination of an object’s position within a room or a factory hall is often needed. Ultra-wideband is a promising technology that in the- ory enables high-accuracy position determination due to its very short-pulsed signals. How- ever, as any other signal-based technology for indoor positioning systems, ultra-wideband signals face the problem that indoor environments are harsh: a lot of reflections at and transmissions through objects make correct position determination difficult. Since changes to the environment (e.g. moving the machines in the factory hall) are impossible, the ques- tion arises where to optimally set up anchors in order to be able to locate objects with high precision. This question is the one this thesis attempts to solve. The approach taken is to first model the environment to be covered in order to then perform a propagation simulation using ray tracing. Afterwards, we use the results of this simulation as inputs to a particle swarm op- timizer that finds the optimal choice of anchor positions based on the importance the user has placed on several different influencing factors. Since a formal verification of global optimality is not possible, extensive measurements and experiments have been conducted in order to increase confidence in the outcomes of the optimization. They show that the software framework created is able to find reasonably good anchor positions. It makes the influence of individual anchors transparent and en- ables users to objectively compare various anchor positions. The program provides a lot of user settings and parameters and therefore offers room for adaptations to changing require- ments. This shows, for the first time, that it is possible to apply Particle Swarm Optimization to the problem of finding the best anchor positions for indoor localization.



more Talks at the RCS