Multipath-assisted Single-anchor Outdoor Positioning in Urban Environments

Sammanfattning: An important aspect of upcoming fifth-generation (5G) cellular communication systems is to improve the accuracy with which user equipments can be positioned. Accurately knowing the position of a user equipment is becoming increasingly important for a wide range of applications, such as automation in industry, drones, and the internet of things. Contrary to how existing techniques for outdoor cellular positioning deal with multipath propagation, in this study the aim is to use, rather than mitigate, the multipath propagation prevalent in dense urban environments. It is investigated whether it is possible to position a user equipment using only a single transmitting base station, by exploiting position-related information in multipath components inherent in the received signal. Two algorithms are developed: one classical point-estimation algorithm using a grid search to find the cost function-minimizing position, and one Bayesian filtering algorithm using a point-mass filter. Both algorithms make use of BEZT, a set of 3D propagation models developed by Ericsson Research, to predict propagation paths. A model of the signal received by a user equipment is formulated for use in the positioning algorithms. In addition to the signal model, the algorithms also require a digital map of the propagation environment. The algorithms are evaluated first on synthetic measurements, generated using BEZT, and then on real-world measurements. For both the synthetic and real-world measurement sets, the Bayesian point-mass filter outperforms the classical algorithm. It is observed how, given synthetic measurements, the algorithms yield better estimates in non-line-of-sight regions than in regions where the user equipment has line-of-sight to the transmitting base station. Unfortunately, these results do not generalize well to the real-world measurements, where, overall, neither algorithm is able to provide reliable and robust position estimates. However, as multipath-assisted positioning, to the best of our knowledge, has not been used for outdoor cellular positioning before, there are plenty of algorithm extensions, modifications, and problem aspects left to be studied - some of which are discussed in the concluding chapters.