End-to-End Positioning and Mapping in Millimeter-Wave Networks : From physical signals to situational awareness

Abstract

Localization and positioning play a significant role in our contemporary connected world. Navigation applications, the tracking of goods, transportation and commuting – all these positioning-based services shape our everyday life. Emerging technologies such as autonomous driving, robotics, internet of things depend on the accurate positioning even more. On the other hand, wireless communication governs a large part of our daily life. The development of fifth generation (5G) and beyond networks provides a convenient infrastructure for positioning by means of wider bandwidths, larger antenna arrays, small cell deployment and transition to the millimeter-wave (mmWave) bands. Communication radio signals are ubiquitous and they become an asset that can be also utilised for positioning and radio sensing, widening the capabilities of users and providing the situational awareness necessary for adoption of emerging technologies. The contributions of this thesis probe the entire cycle of the positioning process: from the signal generation, via channel parameter estimation, to the user tracking and environment mapping. The proposed methods were validated with data from both numerical simulation and real-life measurements. Moreover, the proposed methods, being based on the received signal power measurements, are applicable to a wide population of user equipment working in the mmWave range, including commercial-off-the-shelf (COTS) products.

First, a novel cascaded algorithm for positioning in dense outdoor mmWave networks is proposed. This algorithm, based on the angle of departure (AoD) estimations, is network-centric and allows the communication system to not only know the position of the autonomous users, but also helps to estimate and adjust the uncertainties in the base station (BS) orientation, which may become important in a dynamic network deployment. Second, the indoor mmWave communication system is considered from the point of view of joint communication and sensing (JCAS). Reference signals from both user (uplink) and BS (downlink) are tested for monostatic and bistatic sensing, respectively. In both cases, the user is able to independently map its surroundings in a dynamic way. With uplink monostatic sensing, the user can track the walls of the indoor environment utilizing the dynamic properties of different types of radio reflection. In downlink bistatic sensing, channel parameters such as the AoD and angle of arrival (AoA), are reliably estimated with a novel method based on singular value decomposition (SVD), allowing application of various simultaneous localization and mapping (SLAM) techniques for full user situational awareness.