Publikationen

In der Spezifikation für 5G New Radio (NR) wurden zwei Arten von Codebüchern, nämlich Typ-I und Typ-II, festgelegt. CB des Typs II bieten im Vergleich zu CB des Typs I eine hochauflösende und genauere Darstellung des Ausbreitungskanals. Mehrere Varianten der Typ-IICBs wurden bis zu Rel. 17 des Third Generation Partnership Project (3GPP) NR spezifiziert. Alle Varianten sind jedoch auf Fußgänger oder Szenarien mit geringer Mobilität ausgerichtet. Die aktuellen Typ-II-CBs bieten bei mittleren bis hohen Geschwindigkeiten keine angemessene Leistung, was hauptsächlich auf die derzeitige CSI-Messung und -Berichterstattung auf Basis der Kanalkohärenzzeit zurückzuführen ist, die die Kanalalterungseffekte der Ausbreitungskanäle nicht erfasst. In jüngster Zeit hat die Doppler-basierte Vorcodierung großes Interesse geweckt, da sie in der Lage ist, die Kanalalterungseffekte durch Ausnutzung der Doppler-Informationen der Ausbreitungskanäle zu überwinden. In dieser Masterarbeit wird ein vollständiges Rahmenwerk für Doppler-basierte Vorcodierung auf der Grundlage der aktuellen Typ-II-CBs entwickelt. Verschiedene Schätzverfahren, nämlich der gradientenbasierte Maximum Likelihood Parameterschätzer (RIMAX) und Orthogonal Matching Pursuit (OMP), werden zur Schätzung des Kanals verwendet, und ein auf OMP basierender Schätzalgorithmus mit geringer Komplexität wird vorgeschlagen. Auf der Grundlage des geschätzten Kanals wird die Berechnung der Vorkodierer sowohl an der gNodeB (gNB) als auch an der Benutzerausrüstung (UE) vorgeschlagen. Die vorgeschlagenen Vorkodierer werden in Bezug auf die Leistung und den Rückkopplungs-Overhead mit den Rel. 16 Typ II Vorkodierern verglichen. Darüber hinaus wird die Auswirkung der vorgeschlagenen Vorkodierer in Bezug auf steigende Aktualisierungsraten, Anzahl der Unterbänder und eine Reihe von Überabtastungsfaktoren für die Verzögerungs- und Doppler-Komponenten untersucht. Simulationsergebnisse, die auf der durchschnittlichen gegenseitigen Information basieren, und vorläufige Simulationsergebnisse auf Verbindungsebene zeigen, dass der vorgeschlagene Vorkodierer, der die Doppler- Information ausnutzt, die Rel. 16 Vorkodierer sowohl in Bezug auf die Leistung als auch den Rückkopplungs-Overhead übertrifft.

This chapter is the perfect introduction to get an overview of THz channel sounder technologies. Additionally, all relevant state of the art and references for the field of THz channel sounding are summarized. The aim of the THz sounder chapter is to create a basic understanding of measurement setups and challenges for the measurement of the electromagnetic wave propagation in the THz range. All necessary principles, from generating the transmit signal over different mixing principles to the THz band and the data acquisition, are compact summarized.

For frequencies in the GHz-range, anechoic chambers are usually evaluated using ray-tracing techniques to locate disturbing reflections off the chamber walls. Most approaches reduce this wave-absorber interaction to a specular reflection, although the absorbers may extend over several wavelengths in size and display a rough surface. In order to develop more realistic ray-tracing models, the reflection characteristics of absorbers must be evaluated based on physical wave phenomena. In this paper, a measurement method is proposed which extends the established NRL-arch to measure the angle-dependent reflectivity for non-specular cases. First measurement results of commercial pyramidal absorbers in the frequency range between 1GHz and 10GHz indicate that the assumption of specular reflections is not justified, as power is reflected over a wide angular range with approximately the same intensity. This effect is, to our knowledge, currently not implemented in ray-tracing methods. These results contribute to a better understanding of the properties of RF absorbers to improve the efficiency of their use.

As new cellular technologies are introduced in mobile networks nowadays, more and more small cells and mobile road side units are being deployed to upgrade the performance of the wireless communication network. With regard to the overall human RF exposure, the question arises to what extent the exposure is affected by the deployment of such small cells compared to macro cells as the status quo. In an attempt to answer this important question, this paper proposes a method to measure and evaluate the combined uplink and downlink exposure in terms of the specific absorption rate of a mobile phone user at recently installed small cell locations. The comparison between the small cell and macro cell scenarios revealed a reduction of the total exposure when the user equipment was logged into the investigated LTE small cells, although the downlink exposure increased. The main reason for this encouraging result is the reduction of the uplink transmission power due to the improved link conditions across the small cell. In contrast, the total exposure in a macro cell scenario is dominated by the uplink exposure. In order to minimise the total exposure, a balance must be sought between downlink and uplink exposure.

Background - Transcranial direct current stimulation (tDCS) is a promising tool to enhance therapeutic efforts, for instance, after a stroke. The achieved stimulation effects exhibit high inter-subject variability, primarily driven by perturbations of the induced electric field (EF). Differences are further elevated in the aging brain due to anatomical changes such as atrophy or lesions. Informing tDCS protocols by computer-based, individualized EF simulations is a suggested measure to mitigate this variability. - Objective - While brain anatomy in general and specifically atrophy as well as stroke lesions are deemed influential on the EF in simulation studies, the influence of the uncertainty in the change of the electrical properties of the white matter due to white matter lesions (WMLs) has not been quantified yet. - Methods - A group simulation study with 88 subjects assigned into four groups of increasing lesion load was conducted. Due to the lack of information about the electrical conductivity of WMLs, an uncertainty analysis was employed to quantify the variability in the simulation when choosing an arbitrary conductivity value for the lesioned tissue. - Results - The contribution of WMLs to the EF variance was on average only one tenth to one thousandth of the contribution of the other modeled tissues. While the contribution of the WMLs significantly increased (p≪.01) in subjects exhibiting a high lesion load compared to low lesion load subjects, typically by a factor of 10 and above, the total variance of the EF didnot change with the lesion load. - Conclusion - Our results suggest that WMLs do not perturb the EF globally and can thus be omitted when modeling subjects with low to medium lesion load. However, for high lesion load subjects, the omission of WMLs may yield less robust local EF estimations in the vicinity of the lesioned tissue. Our results contribute to the efforts of accurate modeling of tDCS for treatment planning.