Augmenting cellular networks with sensing and positioning : algorithm development and validation with proof of concepts

Abstract

Since at least Long Term Evolution (LTE), cellular networks have been evolving towards adding functionalities that go beyond communication services in the traditional sense. This thesis focuses on two of the main concepts of this trend: positioning of user equipments by leveraging the transmission of radio signals and the cellular infrastructure, and Integrated Sensing and Communication (ISAC), which is a comparatively new design paradigm that aims at equipping mobile networks with radar-like capabilities. In contrast to positioning, sensing allows to also obtain information about non-transmitting objects, i.e., those that are not part of the network. Many approaches used in the context of ISAC and positioning resort to spectral estimation techniques that exploit the properties of the wireless channel to extract information from incoming radio signals. With this in mind, one of the contributions of this work is an algorithm based on Multiple Signal Classification (MUSIC), which is tailored to the requirements imposed by 5G numerologies and makes joint estimation of range and angle computationally feasible. As positioning in cellular networks can look back on multiple years of research, various concepts and fairly mature solutions are already in place. For this reason, the investigations in this thesis focus on extending existing algorithms that are based on probabilistic time of arrival and angle of arrival localization. In particular, a robust initialization routine is presented that limits outliers and allows finding the global maximum in an efficient manner. The advantages of the solution are demonstrated with measurements from a 5G positioning proof of concept (PoC) in a factory-like environment. ISAC research, on the other hand, is still in a rather early phase. This work deals with monostatic sensing systems utilizing the principles of orthogonal frequency division multiplexing (OFDM) radar, as they presumably represent the first realizations of ISAC deployments in upcoming cellular standards. In that context, an analytical model is adopted to estimate the performance and limiting factors of envisioned ISAC use cases based on to be expected hardware features in the coming years and under consideration of available signal parametrization options. One important ISAC research branch is the suppression of clutter, i.e., of reflections from objects that are not of interest for the sensing task and can thus be regarded as interference. Methods for handling clutter in ISAC applications are extensively discussed in this thesis, which overcome several shortcomings of prior art approaches. The introduced algorithms enable both the acquisition and tracking of information about clutter components, as well as their efficient removal in real-time. The benefits of the techniques are illustrated not only by means of different simulation campaigns, but also using an ISAC PoC for pedestrian tracking experiments in a lab environment.