Konferenzbeiträge

At measurement points (MP) with a line of sight (LOS) to the base station, measurement and extrapolation procedures already exist to determine the maximal exposure in the vicinity of 5G massive MIMO base stations. However, for MPs with non-LOS (NLOS) to the base station, incorrect estimations of the maximal exposure have been found. Based on the theoretical preliminary considerations in part 1 of this paper, the exposure at outdoor NLOS MPs has been determined at 63 MPs in the vicinity of seven Huawei base stations. The gain correction factors previously used by assuming free-space propagation have been compared with the actual gain correction factors derived from the measurements. It is shown that the resulting misestimation and the actual gain correction factors depend on the distance between the MP and the base station. On the basis of this, generalised extrapolation factors are derived for NLOS MPs. Theoretical and numerical investigations are explained in detail in the accompanying paper "part 1".

In the present paper we introduce novel ultra-wideband (UWB) dual-polarized double-directional measurements at sub-THz (300 GHz) in an access point to inside of machine application in an industrial scenario. The results show a sparse spatial/temporal channel with multiple paths from the different metallic objects and their influence on polarization. In addition, different LOS blockage situations were investigated, showing the presence of alternative paths for communications.

To accurately model a communication channel using channel sounding technique, having dual-polarized antenna elements and low mutual coupling between them is required. In this work a Defected Ground Structure (DGS) is proposed to reduce the mutual coupling between two dual polarized mmWave antenna elements in an array structure. The defected ground is made of etched C-shape slots. The presence of the DGS has a strong impact on H-plane coupling where a suppression of 35 dB is achieved, while the reduction of E-plane coupling is about 4 dB. The unwanted effect on the radiation patterns is negligible and the back radiation is not increased.

Organosheets combine the advantages of reinforcement fibers and thermoplastic polymers. By pairing these two materials, composites with outstanding mechanical properties and low densities can be produced. These semi-finished products can be further processed into complex and functionalized components by thermoforming or injection molding. There are a number of different manufacturing processes for continuous fiber reinforced thermoplastics (CFRT), however, most of them require long production times and recurrent melting of the polymer resulting in high energy and manufacturing costs. This study presents a novel extrusion process, that enables a continuous production of reinforced thermoplastic sheets with only one melting step. Due to the high energy efficiency and wide range of processible materials, this process shows a high potential for an economical production of CFRT. To investigate the extrusion process in more detail, the influence of the processing and the flow behavior of the polymer on the impregnation quality and the mechanical properties of the composites were studied. The results showed increasing fiber volume contents with lower polymer viscosities. Furthermore, higher die temperatures and pressures resulted in higher fiber volume contents and thus in higher mechanical properties. The experiments also revealed that a complete impregnation can currently not be achieved without an additional small double belt press due to the line load of the calender, the high viscosity of the melt and the short impregnation time.

Radio based positioning of a user equipment (UE) based on deep learning (DL) methods using channel state information (CSI) fingerprints have shown promising results. DL models are able to capture complex properties embedded in the CSI about a particular environment and map UE’s CSI to the UE’s position. However, the CSI fingerprints and the DL models trained on such fingerprints are highly dependent on a particular propagation environment, which generally limits the transfer of knowledge of the DL models from one environment to another. In this paper, we propose a DL model consisting of two parts: the first part aims to learn environment independent features while the second part combines those features depending on the particular environment. To improve transfer learning, we propose a meta learning scheme for training the first part over multiple environments. We show that for positioning in a new environment, initializing a DL model with the meta learned environment independent function achieves higher UE positioning accuracy compared to regular transfer learning from one environment to the new environment, or compared to training the DL model from scratch with only fingerprints from the new environment. Our proposed scheme is able to create an environment independent function which can embed knowledge from multiple environments and more effectively learn from a new environment.