Thanks to their high bandwidth and, therefore, nanosecond-level time resolution, UWB communications have the potential to perform positioning with centimeter-level accuracy. If we take the example of personal navigation, the user needs to have a device equipped with a UWB transceiver in order to use a UWB positioning systems.
Fortunately, nowadays it is easy to find such devices, as most new-generation smartphones integrate UWB chipsets. On the infrastructure side, we need at least three more UWB devices with fixed and known positions in order to perform localization in a 2D space. By exchanging UWB messages between the user and infrastructure devices, we can localize the user device with centimeter-level accuracy.
UWB-based positioning systems currently face a number of issues. First, most UWB chipsets have a quite high energy consumption (about 3-4 times more than Bluetooth Low Energy), which can prevent their adoption in ultra-low power wearables.
In her thesis, Flueratoru studied the energy consumption of different UWB device architectures and the trade-off between energy efficiency and positioning accuracy. The results show that some architectures can achieve an ultra-low power operation with only a small penalty in the localization accuracy.
A method to eliminate NLOS errors without performing any measurement campaigns
A second pressing issue is that the positioning accuracy of UWB localization systems degrades in non-line-of-sight (NLOS) propagation. This happens, for instance, when there is a wall or another obstacle between the user device and one of the infrastructure devices.
The thesis studies the accuracy of UWB devices under a range of NLOS conditions. Furthermore, Flueratoru proposed a method to eliminate NLOS errors without performing any measurement campaigns, which are currently an expensive and laborious requirement for most NLOS error-mitigation methods.
Finally, the UWB positioning systems of tomorrow will need to scale to a large number of users with little to no performance loss. The thesis presents a novel UWB localization method that can scale to an unlimited number of users, while being robust to NLOS and multipath propagation. This is achieved using a flexible communication scheme for the infrastructure devices.
Flueratoru conducted her research between 2019–2022 under A-WEAR (A network for dynamic wearable applications with privacy constraints), a Marie Sklodowska-Curie European Joint Doctorate/Innovative Training Network, at University Politehnica of Bucharest, Romania, and Tampere University, Finland. A-WEAR received funding through the Horizon 2020 Research and Innovation Programme under the grant No. 813278.
Public defence on 3 July
The doctoral dissertation of M.Sc. Laura Flueratoru in the field of communications titled Robust, Energy-Efficient, and Scalable Indoor Localization with Ultra-Wideband Technology will be publicly examined at the Faculty of Computer Science and Engineering, at University Politehnica of Bucharest, at 10.00 EEST on the 3rd of July 2023, at the PRECIS building (Bvd. Iuliu Maniu 6D, Bucharest, Romania), room PR105. The Opponents will be Doctor Christian Gentner from the German Aerospace Center (DLR) and Professor Radu Tudor Ionescu from the University of Bucharest. The Custos will be Professor Bogdan Dumitrescu from University Politehnica of Bucharest.