4th MC and 4th Technical Meeting - Dubrovnik, Croatia, 23-26 January
TD Number | TD Author | TD Title | TD Abstract | TD WG |
| TD(23)04001 | Ebubekir Memisoglu, Halise Turkmen, Huseyin Arslan | CSI-based NOMA for Integrated Sensing and Communication | The flexible mixed-numerology structure with or-thogonal frequency division multiplexing (OFDM) waveform is a key enabler to meet the requirements of various applications in fifth-generation (5G) and beyond wireless networks. However, the high peak-to-average power ratio (PAPR) is still one of the main drawbacks and conventional numerology selection can provide PAPR reduction only for the time-domain numerology multiplex-ing. In this letter, a novel numerology scheduling (NS) method is proposed to provide PAPR reduction for both time-domain and frequency-domain numerology multiplexing scenarios. For the proposed method, the signal with minimum PAPR is selected from a set of signals with different NS. The simulation results and complexity analysis demonstrate that the proposed method with a linearithmic complexity improves PAPR reduction compared to conventional numerology selection. | WG2 |
| TD(23)04002 | Julien Sarrazin, Guido Valerio | H-Plane-Scanning Multibeam Leaky-Wave Antenna for Wide-Angular-Range AoA Estimation at mm-wave | A fully-metallic periodic leaky-wave antenna well-suited for angles-of-arrival estimation with a reduced frequency bandwidth is presented. The design is based on a leaky rectangular waveguide loaded with corrugations and periodically modulated by longitudinal slots, which produces a frequency beam scanning in the H-plane. By exploiting multiple visible spatial harmonics, multiple beams are radiated for each frequency. This enables the antenna main beam to scan the 180°-angular range with a bandwidth of only 3.7% at 27GHz. Simulations show that using a frequency-domain MUSIC approach, it is possible to estimate angles of arrival of several incoming sources without ambiguity. | WG1,Sub-WG1.1 |
| TD(23)04003 | Francois Quitin, Michel Osée | A wireless transceiver for Control Area Network: proof-of-concept implementation | The Control Area Network (CAN) is a communication standard for cabled bus networks that is widely used in industrial environments. The CAN protocol implements Carrier Sense Multiple Access/Non Destructive Arbitration (CSMA/NDA) protocol for medium access control. This protocol, quite unique to CAN bus communications, preempts collisions by implementing an arbitration process that designates which node will get access to the medium. Since CSMA/NDA requires full-duplex transceivers, it was widely believed that CAN communications could not be implemented with wireless transceivers. In this paper, we demonstrate that it is possible to implement the CSMA/NDA protocol using On-Off Keying (OOK) modulation. We realize a proof-of-concept implementation using off-the-shelf wireless OOK transceivers, and show that our wireless transceivers are fully compatible with CAN controllers available on most microcontroller systems. | WG3,VT3 |
| TD(23)04004 | Markus Ulmschneider, Christian Gentner, Armin Dammann | Learning-Based Fusion of Multipath Assisted Positioning and Fingerprinting | In multipath assisted positioning, multipath components (MPCs) are regarded as line-of-sight (LoS) signals from virtual transmitters. The locations of the physical and the virtual transmitters can be estimated jointly with the user position using simultaneous localization and mapping (SLAM). We have previously introduced such an approach called cooperative Channel-SLAM, where multiple users cooperatively estimate the locations of physical and virtual transmitters. Such schemes typically suffer from a high computational complexity due to expensive signal processing, though. Within this paper, we propose a novel approach that combines multipath assisted positioning with fingerprinting. In the first stage, multiple users estimate their own locations with cooperative Channel-SLAM. With the channel estimates and the estimated user positions from cooperative Channel-SLAM, a deep neural network (DNN) is trained. In the second stage, users can localize themselves making use of the DNN. In our novel approach, the positioning error is in the same order of magnitude as for cooperative Channel-SLAM, while the computational complexity is reduced drastically. | WG2 |
| TD(23)04005 | Michael Walter, Miguel Angel Bellido-Manganell | Differential Geometric Description of Future Mobile-To-Mobile Channels | In this paper we present differential geometric concepts for future mobile-to-mobile channel models. Those concepts enable the reader to generalize the channel description and apply it to different mobile-to-mobile scenarios. The geometric description allows for a non-stationary model, which thus incorporates high mobility scenarios, which are very important in future systems. Hereby, the mathematical concepts are similar to the theory of general relativity. Differential forms and coordinate system transforms are part of a tensor description of the channel. Our channel model for mobile-to-mobile channels is based on two important mathematical concepts: prolate spheroidal coordinates and differential forms. Furthermore, both the coordinate transform and the differential forms can be uniformly described by the tensor theory. Thus, we use co- and contravariant tensors to express both the gradient of the Doppler frequency in prolate spheroidal coordinates and differential forms, which are used to calculate the scattering area, in a uniform way. | WG1 |
| TD(23)04006 | Manuel Castillo-Cara, Reewos Talla-Chumpitaz, Luis Orozco-Barbosa and Raúl García-Castro | From Tidy Data into Synthetic Image: A novel mechanism for indoor localization | The present work develops a system for indoor localisation prediction using Bluetooth-based fingerprinting using Convolutional Neural Networks (CNN). For this purpose, a novel technique was developed that simulates the diffusion behaviour of the wireless signal by transforming tidy data into images. For this transformation, we implemented the technique used in painting known as blurring, simulating the diffusion of the signal spectrum. Our proposal also includes the use and a comparative analysis of two dimensional reduction algorithms, PCA and t-SNE. Finally, an evolutionary algorithm was implemented to configure and optimise our solution with the combination of different transmission power levels. The results reported in this work present an accuracy of close to 94%, which clearly shows the great potential of this novel technique in the development of more accurate indoor localisation systems | WG2,HA1 |
| TD(23)04007 | Paul Unterhuber, Michael Walter, Thomas Kürner | Parametrization and Validation of the Geometry-Based Stochastic Channel Model for Train-to-Train Communication | The future of railway transport aims at more flexible train composition with virtually coupled train sets (VCTSs), automatic train operation, and autonomously driving trains. The train-to-train (T2T) communication is one major technology which is enabling those future railway concepts. Hence, the development of the next generation of communication standards takes into account railway communication. The development of standards and equipment for wireless communication requires a profound knowledge of the propagation conditions in railway environments and for moving trains. For safety critical applications the time-variant behavior of the propagation channel has to be considered. Therefore, we propose a geometry-based stochastic channel model (GSCM) and show the parametrization for an open field environment. The proposed and parametrized GSCM enables the development and testing of communication standards for T2T communication. | WG1,VT2 |
| TD(23)04008 | Xiping Wang, Ke Guan Danping He, Zhao Zhang, Haoyang Zhang, Jianwu Dou, Ruiqi (Richie) Liu and Zhangdui Zhong | Super Resolution of Wireless Channel Characteristics: A Multi-Task Learning Model | —Channel modeling has always been the core part of the design and development of communication system, especially in 6G era. Traditional approaches like stochastic channel modeling and ray-tracing (RT) based channel modeling depend heavily on measurement data or simulation, which are usually expensive and time-consuming. In this paper, we propose a super resolution (SR) model for recovering high-resolution channel characteristics from sparse sampling data. The model is based on multi-task learning (MTL) convolutional neural networks (CNN) with attention mechanism and residual connection. Experiments demonstrate that the proposed MTL SR model could achieve fairly good performances in terms of mean absolute error and standard deviation of error. The advantages of the proposed model are demonstrated in comparisons with other state-of-theart deep learning models and visualization of SR performances. The ablation study also proved the necessity of data augmentation and the techniques in model design. The good generalization ability of the proposed MTL SR model leads to discussions of potential applications. The contribution of this paper could be helpful in channel modeling, network optimization, positioning, and other wireless channel characteristics related work by largely reducing the workload of simulation or measurement. | WG1,WG2,VT2,HA1,HA3 |
| TD(23)04009 | Abigail Elcock and Alister Burr | Optimum linear fronthaul quantization for Open-RAN systems | The Open Radio Access Network architecture assumes that the physical layer may be disaggregated between the Radio Unit (RU), which normally includes RF functions, the FFT/IFFT and other lower physical layer functions, and the Distributed Unit (DU) which implements the higher physical layer functions, including modulation/demodulation and FEC coding/decoding. These components are connected via fronthaul connections, which carry digitised signals. To minimise the load on these fronthaul connections coarser quantization (referred to as fronthaul compression) is applied, which introduces quantization distortion and hence may affect the performance of the physical layer, especially on the uplink with soft decision decoding in the DU. In this paper we use direct simulation of the physical layer to determine optimum quantization intervals and the corresponding physical layer performance for a range of standard 5G-NR modulation and coding schemes. We show that depending on the modulation order much coarser quantization may be used than is envisaged in the fronthaul compression schemes included in current O-RAN standards, with negligible performance loss. We also compare with optimum quantization thresholds and physical layer performance obtained using an information theoretic approach, and discuss further work required to generalise the results to more realistic scenarios. | WG2 |
| TD(23)04010 | Agnieszka Czapiewska, Andrzej Luksza, Ryszard Studanski, Andrzej Zak | Analysis of Impulse Responses Measured in Motion in a Towing Tank | The growing interest in developing autonomous underwater vehicles (AUVs) and creating underwater sensor networks (USNs) has led to a need for communication tools in underwater environments. For obvious reasons, wireless means of communication are the most desirable. However, conducting research in real conditions is troublesome and costly. Moreover, as hydroacoustic propagation conditions change very significantly, even during the day, the assessment of proposed underwater wireless communication methods is very difficult. Therefore, in the literature, there are considered simulators based on real measurements of underwater acoustic (UWA) channels. However, these simulators make an assumption that, during the transmission of elementary signals, the impulse response does not change. In this article, the authors present the results of the measurements realized in a towing tank where the transmitter could move with a precisely set velocity and show that the analyzed channel was nonstationary, even during the time of the transmission of a single chirp signal. The article presents an evaluation method of channel stationarity at the time of the chirp transmission, which is the novelty of the proposed work. There is also an analysis of the impulse responses measured in motion in a towing tank. The estimated impulse responses are attached to this article as supplementary materials and are the main contribution. | WG1 |
| TD(23)04012 | Piotr Rajchowski, Krzysztof K. Cwalina, Jarosław Sadowski | Analyzing the Influence of the Radio Channel on Precision of Position Estimation of the user Terminal Using the NB-IoT Radio Interface | In the TD the method of user terminal position estimation using the occasional signals transmitted in the down-link of the NB-IoT interface was presented, including the analysis of influence of the radio channel on precision of estimated positions. The practical ability to estimate the position in the test area was verified using the reference signals with a variable signal-to-noise ratio and different radio channel profiles. Additionally, the sample method is presented that allows to increase the precision of the terminal position estimation without the need of increasing the sampling frequency of the radio signal. | WG2,VT3,VT4 |
| TD(23)04013 | Kai Mao, Qiuming Zhu, Yang Miao | A UAV-Aided Real-Time Channel Sounder for Highly Dynamic A2G Scenarios | With the rapid development and broad applications of unmanned aerial vehicle (UAV) based wireless stations in the sky, fundamental understanding and characterization of the realistic A2G communication link properties are crucial. In this paper, a UAV-aided channel sounder with a real-time processing hardware system is developed for highly dynamic and non-stationary A2G channel measurements. In the hardware system, several real-time hardware processing algorithms for raw channel data, i.e., channel impulse response (CIR) extraction, system response elimination (SRE), sampling time offsets compensation (STOC), and adaptive multi-path component (MPC) recognition are developed and implemented on a single field- programmable gate array (FPGA) chip. In this way, the required storage size of channel data and the processing time for one slice of CIR is greatly decreased, which can meet the requirement of non-stationary A2G channel measurement with a high sampling rate and long-time measurement. A commercial channel emulator is used to reproduce controllable channels and verify the performance of developed channel sounder. Finally, the developed channel sounder is applied to carry out A2G measurement campaigns at 3.5 GHz in a campus scenario. The channel sounder can be used to capture the non-stationary A2G channel characteristics for the system design and algorithm optimization of A2G communications. | WG1 |
| TD(23)04014 | Dheeraj Raja Kumar, Carles Antón-Haro and Xavier Mestre | Neural Network-based MIMO Channel Estimation: Shallow vs. Deep Architectures | In this paper, we investigate neural network-based channel estimation strategies for point-to-point multi-input multi-output (MIMO) systems. In an attempt to keep computational complexity low, we restrict ourselves to shallow architectures with a single hidden layer. Specifically, we consider (i) fully-connected feedforward neural networks; and (ii) 1D/2D convolutional neural networks. The analysis includes an assessment of the estimation error performance, along with the computational complexity associated to the training and inference phases. Several benchmarks are considered, such as the conventional least squares or (linear) MMSE estimators, and other deep neural network architectures from the literature. | WG2 |
| TD(23)04015 | Manuel M. Ferreira, Filipe D. Cardoso, Sławomir J. Ambroziak and Luis M. Correia | Bandwidth Dependence of the Propagation Channel in Circular Metallic BAN Environments | In this paper, the bandwidth dependence of the propagation channel for Body Area Networks (BANs) in circular metallic environments is addressed and models are proposed to evaluate the appropriate short-term fading margins that should be considered as a function of the system bandwidth. The deployment of BANs in metallic indoor environments, such as ships, factories, warehouses and other similar environments, involves additional challenges compared to other indoor environments due to the specific propagation effects in this type of environments (i.e., with strong reflections), giving the motivation for this work. Bandwidth dependent values of delay spread are also presented and discussed. It is observed that for a system bandwidth up to 100 MHz the fading depth is composed of three main stages, with the transitions between these stages being associated with the system ability to discriminate arriving rays at the receiver. Average values of fading depth are 16.4, 13.2 and 11.0 dB for stages 1, 2 and 3, respectively, the difference between consecutive stages ranging between 2.2 to 3.2 dB. For a system bandwidth larger than 100 MHz, the fading depth decreases with an increasing system bandwidth, with an average decay rate close to 3 dB per 100 MHz bandwidth, being about 2 dB for system bandwidths above 400 MHz. | WG1 |
| TD(23)04016 | Milica Lekić, Gordana Gardašević, Milan Mlađen | EXPERIMENTAL EVALUATION OF MULTI-PHY 6TISCH NETWORK | The architectural design of Wireless Sensor Networks (WSNs) for the Industrial Internet of Things (IIoT) applications requires a careful planning and selection of proper operational strategy. Harmonization of standards is crucial to ensure easier certification and commercialization of IIoT solutions. The ongoing research activities are directed toward designing agile, reliable, and secure transmission technologies and protocols. Recently, Time Slotted Channel Hopping (TSCH) standardization bodies have started to consider the support for multiple physical layers thus accommodating a wide range of applications. This paper presents the results of the extensive experimental measurement campaign to study the performance of the 6TiSCH (IPv6 over the TSCH mode of IEEE 802.15.4e) network while supporting multiple physical layers (PHYs). For measurement purposes, all experiments were performed on OpenMote-B hardware. These devices are equipped with an Atmel AT86RF215 dual radio transceiver implementing the IEEE 802.15.4g. The performance evaluation is provided for the following metrics: network formation time, packet delivery ratio (PDR), latency, and duty cycle. Results are encouraging, particularly in terms of high PDR for all tested PHYs. Performance evaluation indicates the importance of proper node positioning, link quality estimation and careful selection of network parameters. Moreover, collected experimental results create a dataset that provides insights into the tested PHYs performance and their potential for the indoor 6TiSCH networking. | VT3 |
| TD(23)04017 | Zhao Zhang, Danping He, Xiping Wang, Ke Guan, Zhangdui Zhong, Jianwu Dou, Stephen Wang | A Ray-tracing and Deep Learning Fusion Super-resolution Modeling Method for Wireless Mobile Channel | Mobile channel modeling has always been the core part for design, deployment and optimization of communication system, especially in 5G and beyond era. Deterministic channel modeling could precisely achieve mobile channel description, however with defects of equipment and time consuming. In this paper, we proposed a novel super resolution (SR) model for cluster characteristics prediction. The model is based on deep neural networks with residual connection. A series of simulations at 3.5 GHz are conducted by a three-dimensional ray tracing (RT) simulator in diverse scenarios. Cluster characteristics are extracted and corresponding data sets are constructed to train the model. Experiments demonstrate that the proposed SR approach could achieve better power and cluster location prediction performance than traditional interpolation method. Channel impulse response (CIR) is reconstructed based on cluster characteristics, which could match well with the multipath component (MPC). The proposed method can be used to efficiently and accurately generate big data of mobile channel, which significantly reduces the computation time of RT-only. | WG1 |
| TD(23)04018 | Haofan Yi, Ke Guan, P. Takis Mathiopoulos, Pengxiang Xie, Danping He, Jianwu Dou, Zhangdui Zhong | Full-Wave Simulation and Scattering Modeling for Teraherz Communications | This paper presents a comprehensive study for analyzing and modeling the scattering phenomenon of THz waves bouncing on rough surfaces. Firstly, a generic parametric methodology is proposed to accurately model their operational characteristics focusing on their root-mean-square (RMS) heights and correlation lengths. A Monte Carlo-based approach is developed for the efficient and accurate software realization of such rough surfaces. Then, full-wave simulations are used to test at 300 GHz the far-field behavior of scattering waves on 30 distinct surfaces having 6 RMS heights and 5 correlation lengths. By employing the directive scattering (DS) model, the variation of the scattering coefficient, S, and its equivalent roughness, alpha_R, are studied for the scattered signal amplitude aspects. Furthermore, the statistical characteristics of the phase and polarization are obtained and compared with those already available in existing 3GPP channel model standards which are valid for lower frequency bands. These comparisons have shown that although the phases follow the normal distribution, which aligns with equivalent 3GPP channel models, the crosspolarization ratios (XPRs) follow a different distribution, namely the logistic distribution. This is a very important difference which could have a great impact on the standardization activities of the current 3GPP-related working groups. Finally, the proposed scattering model at the THz band is detailed with algorithmically implemented steps. Compared with the classical two-dimensional DS model charactering only the amplitude of the scattered signal, the proposed three-dimensional scattering model can be accurately modeled by characterizing the most important parameters of the scattered signals, namely, amplitude, phase, and polarization at the THz band. A further advantage of the proposed scattering model is that it can be easily integrated with a ray-tracing (RT) scheme thus allowing precise scattering and channel modeling for new applications, such as holographic radios and multi-user multiple-input multiple-output (MU-MIMO) systems. | WG1 |
| TD(23)04019 | Greta Vallero, Michela Meo | Modelling Solar Powered UAV-BS for 5G and Beyond | Unmanned Aerial Vehicles equipped with Base Stations (UAV-BSs) are considered an effective solution to dynamically provide additional capacity in Radio Access Networks (RANs), in case of network congestion or emergency situations. To face the problem of the poor energy availability provided by on-board batteries, UAV-BSs can be equipped with Photovoltaic (PV) panels. To investigate and understand the complex interworking between traffic needs and energy availability, in this paper, we propose a model of a PV-panel-powered Long Term Evolution (LTE) Multi User Multiple Input-Multiple Output (MU-MIMO) UAV-BS, using a discretized representation of the energy in terms of Energy Packets (EPs). The model highlights the different operation regions under which the traffic demand can be satisfied for given energy production levels. Results for winter and summer seasons that take into account daily traffic and energy production variability are shown and can be used to properly dimension the UAV-BS power supply system. | WG3 |
| TD(23)04020 | Hamid Taramit, Luis Orozco-Barbosa, Abdelkrim Haqiq, José Jaime Camacho Escoto, Javier Gomez | Adaptive Channel Allocation in non-Saturated Rayleigh Fading Wi-Fi HaLow Networks | This paper considers a non-saturated IEEE802.11ah-based network where stations are distributed around the Access Point (AP) and operate under a Rayleigh-fading channel with capture. We develop an analytical framework based on a two-level renewal process to model the contention within a group of stations and derive the Average Service Time (AST), presenting the proper size of the RAW slot that ensures successful delivery of all packets without extra waste of channel time. We also evaluate the Packet Delivery Ratio (PDR) and channel usage for a pre-allocated RAW slot in terms of designated stations to prove the effectiveness of our proposal. Our analytical framework is validated via extensive MATLAB simulations and can be applied to alternative communication technologies addressing dense scenarios and integrating periodic channel reservations. | WG3,VT4 |
| TD(23)04021 | Jaeyoung Park, Ali Al-Ameri, Juan Sanchez, Xuesong Cai, and Fredrik Tufvesson | A Hybrid Antenna Switching Scheme for Dynamic Channel Sounding | Channel sounding is essential for the development of radio systems. One flexible strategy is the switched-array-based channel sounding, where antenna elements are activated at different time instants to measure the channel spatial characteristics. Although its hardware complexity is decreased due to fewer radio-frequency (RF) chains, sequentially switching the antenna elements can result in aliasing in the joint estimation of angles and Doppler frequencies of multipath components (MPCs). Therefore, random switching has been proposed to mitigate such aliasing and subsequently increase estimation accuracy in both angular and Doppler frequency domains. Nevertheless, the increased resolution of Doppler frequencies could cause additional post-processing complexity of parameter estimation, which is relevant when the Doppler frequencies are not of interest, e.g., for spatial channel modeling. This paper proposes an improved hybrid sequential and random switching scheme. The primary purpose is to maintain the estimation accuracy of angles of MPCs while decreasing the resolution of Doppler frequencies for minimized post-processing complexity. A simulated-annealing algorithm is exploited to obtain the optimized switching sequence. The effectiveness of the proposed scheme is also demonstrated with a realistic antenna array. | WG1,Sub-WG1.1 |
| TD(23)04022 | Aleš Simončič, Tomaž Javornik, Andrej Hrovat | Uniform circular antenna array for direction of arrival estimation | This study presents the design and performance evaluation of a uniform circular antenna array (UCAA) with 12 vertical monopole antennas and RF switches to select the active antenna for direction of arrival (DoA) estimation applications. The use of the UCAA configuration and monopole antennas is justified. The antenna configuration was optimised through simulations for a varying number of antennas and radius lengths by simulating a signal impinging on the antenna array at an arbitrary angle and determining that angle using the DoA MUSIC estimation algorithm. The design process of the RF switching matrix and the RF part of the design are described in detail. The frequency tuning of the antennas also considers the coupling of the antennas with parasitic antennas and a comparison is made with and without parasitic antennas. The performance of the antenna array was evaluated by (1) simulations that analyses non-idealities such as the proximity to the edge of the antenna array and the influence of neighbouring antennas on the antenna performance, and (2) measurements in a semi-controlled environment. We achieved an accuracy of less than 1° for the DoA estimate using the Bluetooth Low Energy (BLE) angle-of-arrival feature added in version 5.1.The results of the simulations also show that antenna coupling can cause nulls in the radiation pattern, which affects the DoA determination. | WG2,Sub-WG2 |
| TD(23)04023 | Grega Morano, Ke Guan, Andrej Hrovat, Tomaž Javornik | Phase-based ranging within TSCH communication | Localization and sensing inputs are enabling new and improved Internet of Things (IoT) applications and have been studied extensively over the past decade. Unfortunately, many solutions focus primarily on improving localization performance, which in turn reduces communication capabilities. Driven by the Integrated Sensing and Communication (ISAC) aspects, we investigate how localization functionality can be seamlessly integrated into a IEEE 802.15.4 TSCH communication protocol. We present two new methods of phase-based ranging that estimate the distance between two devices with each transmitted data packet. We further analyze the effects of introduced changes on communication, their power consumption, and ranging operability. We improve the current state-of-the-art algorithm of phase-based distance estimation by reducing the number of phase samples required, without reducing its accuracy and sensitivity. | Sub-WG2 |
| TD(23)04024 | Werner G.~Teich, Thanawat Thiasiriphet | Combining OFDM-MFSK and Chirp Spread Spectrum Modulation for Applications in Low-Power Wide-Area Networks | For the unlicensed frequency bands, long-range wide-area networks (LoRaWAN) is a widespread solution for low-power wide-area networks. The physical layer of LoRaWAN is long range (LoRa) which uses a combination of chirp spread spectrum (CSS) and M-ary frequency-shift keying (M-FSK) for transmission. LoRa is known to be highly robust and well-suited for a low-complexity implementation, but the weak aspect is its low spectral efficiency. Orthogonal frequency division multiplexing M-FSK (OFDM-MFSK) is a technique which combines M-FSK with OFDM. The bandwidth is divided into several parallel M-FSK transmissions, i.e., more than one sub-carrier is active at a time. The reception of OFDM-MFSK can be done non-coherently. Therefore, it holds many similar design principles as LoRa. We propose the combination of OFDM-MFSK and CSS in this paper. The weakness and strength of the two techniques complement each other well. LoRa and OFDM-MFSK with CSS have a similar receiver sensitivity and show comparable performance with respect to power and bandwidth efficiency. However, compared to LoRa, OFDM-MFSK with CSS gives a larger flexibility with respect to the choice of data rates and it is more robust to impulsive noise. Furthermore, simulation results show, that OFDM-MFSK with CSS provides substantial performance improvements compared to unspread OFDM-MFSK, in both, multipath-induced frequency selective channel and for channel with strong narrowband interference. | WG2 |
| TD(23)04027 | Christian Gentner, Martin Schmidhammer | Radio Localization and Sensing in Low-Cost Ultra-Wideband Networks | Device-free localization (DFL) is a prominent example of radio sensing and radio frequency (RF)-based passive localization. RF-based passive localization approaches determine the position of a non-cooperative user based on the user’s impact on radio propagation. In this regard, DFL systems measure user-induced changes in the properties of the received RF signals. With multipath-enhanced device-free localization (MDFL), we introduced a novel passive localization approach which is taking the advantage of user-induced fading in the multipath signals, i.e. reflected and scattered signals. In this work, we realize an MDFL system using low-cost ultra-wideband (UWB) devices. Specifically, we use the Qorvo (DecaWave) DW1000 module, for which we describe how to access the channel impulse response in detail. We additionally provide an overview of the required signal processing for MDFL and introduce a possible sequential Bayesian approach. Based on a fully-meshed network of UWB devices, we evaluate the localization performance of both DFL and MDFL for an indoor scenario. Thereby, MDFL is shown to clearly outperform DFL in terms of robustness and accuracy. Furthermore, we use the measurement setup for active localization, i.e., for localizing an UWB device carried by the user, allowing for a direct comparison between active and passive localization. Achieving a sub-decimeter accuracy, we reveal the applicability of active localization as an inherent ground truth system required for the initialization of MDFL systems. | WG2 |
| TD(23)04028 | Manuel M. Ferreira, Filipe D. Cardoso, Sławomir J. Ambroziak, Mariella Särestöniemi, Kenan Turbic, Luís M. Correia | Depolarisation Model for a BAN Indoor Scenario | This paper presents an analysis of depolarisation in Body Area Networks for Body-to-Infrastructure communications based on a measurement campaign in the 5.8 GHz band in an indoor environment. The measurements were performed with an off-body antenna transmitting linearly polarised signals and with dual-polarised receiving antennas worn on the body by the user. A normal distribution with a mean of 2.0 dB and a standard deviation of 4.3 dB is best fit for modelling cross-polarisation discrimination. The average correlation between the signals received by the orthogonally polarised antennas is less than 0.5, which shows that polarisation diversity can be used. A model is proposed for the average value of the standard deviation of the cross-polarisation discrimination ratio as a function of the transmitted polarisation, the mobility of the users and the link dynamics. | WG1 |
| TD(23)04030 | Teodora Kocevska, Tomaž Javornik, Aleš Švigelj, Ke Guan, Aleksandra Rashkovska, Andrej Hrovat | Generalization of machine learning models for CIR-based materials identification in indoor radio environment | The identification of the materials present within the indoor propagation environment is crucial in providing a comprehensive description of the indoor space. This is of particular importance for the purpose of augmenting the digital twin of the indoor environment with properties of the space that affect and guide the propagation. The detailed understanding of the indoor surrounding of a wireless link provides a significant potential for the wireless system design and performance optimization, as well as for a range of emerging application scenarios in various domains. In this work, we investigate the use of machine learning models constructed from labeled channel impulse response (CIR) data of traced paths for the purpose of predicting the materials present in the surroundings of a wireless link. To date, the literature has not addressed the issue of how the diversity of link positions and propagation environments included in the training data set impacts the generalization performance of the resulting models. Our focus is on the construction of models using varying link positions and environments, and the investigation of their generalization capabilities with respect to unseen communication layouts and room sizes. The results of our study confirm that including a wide diversity of links and spaces in the training process will improve the generalization performance of the models. | Sub-WG2,VT4 |
| TD(23)04031 | Petr Hron, Jan Sýkora, Laurent Clavier | Linear precoding for asynchronous WPNC 2-source hierarchical multiple access channel | Abstract— This paper addresses the problem of asynchronous reception in a 2-source Hierarchical MAC (H-MAC) in the context of a Wireless Physical Layer Network Coded (WPNC) system with a Hierarchical Decode and Forward (HDF) relaying strategy. This work is motivated by the difficulty of precise synchronization among different sources of an H-MAC stage. Because of the non-orthogonal nature of the H-MAC even when the individual delays are known to the receiver a simple compensation is not possible. Compensation at the transmitters would require a channel state feedback channel which is not always available. In some scenarios with varying distances between individual nodes even the knowledge of the delays to the transmitter, a simple time base adjustment would not solve the problem w.r.t. all participating receivers. In this work, we assume delays of a fraction of the symbol time duration and propose a delay-independent transmitter-side linear precoding together with symbol-spaced and fractionally-spaced processing at the receiver. For each approach, we derive an equivalent channel model and analyze its properties in the case of an root-raised-cosine (RRC) modulation pulse. The proposed techniques are numerically evaluated in terms of Hierarchical Bit Error Rate (H-BER). We show that in some cases the asynchronous reception can lead to a performance boost. | WG2 |
| TD(23)04032 | Carsten Smeenk, Zhixiang Zhao | Radio Resource Allocation for Radar Sensing in ISAC systems with Deep Reinforcement Learning | In integrated sensing and communication (ISAC) systems, the radar and communication functionality share the same channel and radio resources. For co-existing functionalities, dedicated signals for radar sensing and communication are designed from the same resource pool without interference. The shared resource pool requires allocation strategies to satisfy radar and communication performance simultaneously. A deep reinforcement learning (DRL) approach is proposed in this TD to adjust the dedicated radar signal. The goal is to guarantee reliable radar detection in a multi-target highway scenario and save radio resources for communication purposes at the same time. The DRL agent observes the output of a multi-target tracking system to predict the optimal radio resources. The training is based on simulated data and an implemented radar signal processing chain. | Sub-WG2 |
| TD(23)04034 | Nikita Lyamin, Fredrik Tufvesson, Aleksei Fedorov | A realistic V2X real-time simulation with spatially consistent channel modelling | We present a real-time SIVERT system-level Vehicle-to-Everything (V2X) simulation framework. It is build around the bi-directional coupling of the VENERIS project implemented in Unity3D and Network Simulator 3. The COST IRACON Geometry-based Stochastic Channel Model (GSCM) approach is adopted for spatially consistent channel modelling in a rich multi-path propagation environment. Antenna system can be included via measured or simulated radiation pattern represented via its Effective Aperture Distribution Function (EADF). The simulation framework runs in a real-time and allows evaluating V2X performance at the early solution design stages covering all C-ITS layers (PHY throughout application layers), including Human-In-The-Loop capabilities to evaluate UX for applications where V2X provides informative support to a drivers. | WG1,WG3,VT2,VT4 |
| TD(23)04035 | Hadi Alidoustaghdam, Min Chen, Ben Willetts, Kai Mao, Andre Kokkeler, Yang Miao | Human Shape and Vital Sign Estimation by Combined CNN and FMCW Radar | Human posture is the key to infer human activities in health and safety monitoring systems, and hence is crucial for the timely intervention towards alarming human activities. Radio technology enabled contact-free human posture and vital sign estimation is promising due to the nonintrusive nature and possibility of operating regardless of the lighting condition. Radio systems at millimeter-wave (mmWave) frequencies advan- tageously bring large bandwidth, multi-antenna array and beam steering capability. However, point clouds of human which are obtained by mmWave radar, and utilized for posture estimation, are likely to be sparse and incomplete, depending on the distance and surrounding environment between the human target and radar. Additionally, human’s random body movements deterio- rate the process of estimating breathing and heart rates, therefore a narrow radar beam toward the chest is demanded for vital sign estimation relies on a proper cross-range resolution of hardware in use. In this paper, we propose to exploit Convolutional Neural Networks (CNN) to efficiently process the imperfect human form point cloud for body part representation and posture estimation, then utilize this information for vital sign estimation. The 3D human point cloud is projected in two vertical planes and a CNN framework is employed to estimate the key point coordinates of different body parts, while Augmentative Ellipse Fitting Algorithm (AEFA) applied on the 2D RGB image (captured by camera that is placed at the same location as radar) provides the ground truth of human’s posture. Finally, the identified key points are stitched together to recognize the human posture. In summary of our proposed pipeline, initially CNN is trained based on the frame-by-frame point clouds of human for posture estimation, and then the vital signs are extracted through beamforming toward the human chest. The numerical results show that this spatial filtering improves the estimation of the vital signs in regard to lowering the level of side harmonics and detecting the harmonics of vital signs efficiently. | VT1 |
| TD(23)04036 | Siwei Zhang, Emanuel Staudinger, Robert Pöhlmann, Fabio Broghammer, Armin Dammann | Radio-based Swarm Navigation: From Theory to Space-Analog Missions on Etna Volcano | A swarm of robots can rapidly explore challenging extraterrestrial environment, such as lunar caves or Martian canyons. Compared to a single robot, a swarm can make simultaneous observations at different locations and avoids a single point of failure, which leads to a paradigm shift in exploration missions. For the success of an autonomous swarm mission, reliable communication and navigation within the swarm is mandatory. We design a self-organized swarm navigation network. This network is organized in a decentralized fashion, adapting its topology and formation to improve the position and orientation estimation of individual robots. In July 2022, we successfully demonstrated our self-organized swarm navigation technology in a space-analog surface exploration mission on volcano Etna, Sicily, Italy. In this mission, a group of robots performed coordinated exploration, while their positions and orientations are precisely estimated with our technology. This paper provides insight on our methods, system design and first results from the analog mission, which sheds light on the usage of radio-signal for joint communication and navigation in a robotic swarm for future exploration missions. | WG2 |
| TD(23)04037 | Kamil Szczeszek and Pawel Kulakowski | AI-aided Base Stations Placement in 5G Mobile Networks | Mobile network greenfield deployments are complicated endeavors where locations of base station to be installed should be chosen carefully to guarantee the best possible network performance. In this paper, we analyze how artificial intelligence algorithms could help in this task. We consider three heuristic approaches: simulated annealing, genetic algorithm and integer linear programming. We adapt them for this specific problem and optimize the base station locations focusing on obtaining the highest possible radio signal coverage. We verify the algorithms performance using a ray tracing implementation to calculate the signal coverage on six maps representing different urban areas. The results show that integer linear programming is the most promising solution, both regarding the calculation time and the obtained signal coverage. | WG3 |
| TD(23)04038 | CheChia Kang, Xin Du, and Jun-ichi Takada | Synchronized Dynamic Channel Sounder and Posture Capture for Millimeter Wave Radio Channel Suffered from Human Body Shadowing | Millimeter wave (mmW) band has been deployed in the fifth generation (5G) mobile communication system to realize the enhanced mobile broadband (eMBB) service. Systems at mmW points the main beam of the base station to the line-of-sight (LoS) toward the mobile stations but suffer from a deep fading due to human blockage. To predict the impact due to human blockage, a synchronized measurement system including dynamic channel sounding and motion capture is needed. This paper proposes the synchronization method between 2 different types of instruments by using a bottom trigger as the reference of time. For validation, the measurement of dynamic channel and motion capture, and the prediction of the channel gain based on uniform theory of diffraction (UTD) were done. The results show that the proposal matching the time stamps well. The proposal provides a reliable synchronization of channel sounder and motion capture for developing dynamic channel model. | WG1,Sub-WG1.1 |
| TD(23)04039 | Gurjot Singh Bhatia, Yoann Corre | Tuning of Ray-Based Channel Model for 5G Indoor Industrial Scenarios | This article presents an attempt to produce deterministic channel models for 5G industrial internet-of-things (IIoT) scenarios. Ray-tracing channel emulation can capture many of the specific properties of a propagation scenario, which is greatly beneficial when facing the large variety of industrial environments and deployment setups. But the complexity of the environment, which is composed of many metallic objects of various sizes and shapes, pushes the ray-tracing tool to its limits. In particular, the scattering or diffusion phenomena can bring significant components. Thus, in this article, the Volcano ray-tracing channel simulation is benchmarked against field measurements found in the literature at two frequencies relevant for 5G industrial networks: 3.7GHz (mid-band) and 28GHz (millimeter-wave (mmWave) band). Both specular and diffuse scattering contributions are calculated; then the ray-tracing data is compared to measured large-scale parameters, such as the power delay profile (PDP), the cumulative distribution function (CDF) of delay spreads, both in line-of-sight (LoS) and non-LoS situations (NLoS) and the IIoT channel properties are further explored. | WG1,VT3 |
| TD(23)04040 | Yann Maret, Jean-Frederic Wagen, Franck Legendre, Mohsin Raza, Nik Bessis | GNN based link predictions to improve the resource allocation in realistic real time emulations of MANETs | Mobile Ad-hoc Networks (MANETs) are required for search and rescue applications where classic infrastructures such as cellular do not offer coverage in remote places. MANETs are composed of radio mobiles that are self reconfigurable and adaptative to the dynamic of the topology. MANETs cope with the mobility of the scenario using a routing protocol. The distributed Optimized Link State Routing (OLSRd2) protocol is considered. OLSRd2 computes routes from the source to destination nodes on each node, leading to a node view graph. Node view graphs do not capture all available links between nodes in the network. GNN-Link Threshold (GNN-LT), a distributed deep learning agent, is proposed to learn the global graph from partial graphs by predicting remaining links. GNN-LT is trained to classify remaining links on realistic mobile scenarios with 24 nodes. Distributing a global graph on nodes allows to compute a more efficient Time Division Multiple Access (TDMA) schedule than the classical round robin schedule. A cooperative and locally compatible suboptimal schedule is derived from the global graph on each node. A traffic scenario generator was developed to test and assess the performance of the GNN-LT agent. GNN-LT is evaluated on the open source Anglova.net MANET scenario. The performance metrics are the completion ratio, the round trip time and the throughput. The results show that GNN-LT can leverage the OLSR node view graphs to predict, on each node, an intermediate graph that captures essential information of the global topology. The derived schedule leads to higher throughput from 3.6kbps to 4kbps when adapting the schedule compared to the classical approach using a round robin schedule. The delay is reduced from 0.7s to 0.6s and the completion ratio is kept to 100%. | WG3 |
| TD(23)04041 | Youssef AGRAM, Jianqiao CHENG, François QUITIN | Direction-of-Arrival estimation using virtual dual-antenna receivers : algorithms and controlled wireless experiments | Localization of radio-frequency (RF) transmitters can be achieved using virtual antenna arrays (VAA) through direction-of-arrival (DoA) estimation. The method relies on a mobile, single-antenna receiver that captures successive messages from a transmitter, thereby emulating a multi-antenna array. However, two main challenges emerge from this technique: 1) Successive positions and orientations of antennas have to be determined, meeting spatial Nyquist criterion; 2) the local oscillator frequency offset (LOFO) between transmitter and receiver adds a drifting phase component to the received signal on each antenna of the array. Previous papers have shown the unability to recover the DoA for rectilinear trajectories using single-antenna array in presence of drifting LOFO. Nonetheless, it has been theoretically proven that integration of a second antenna allows us to return to an observable system. In this paper, we extend the algorithm to dual-antenna receivers, which allows to recover the azimuth. A linear model of the drifting LOFO is also used, allowing for low-quality, drifting Local Oscillators (LOs). Wireless and controlled experimental results from a software-defined radio testbench are presented to corroborate previous theoretical results. Increased performance is noted when moving from a single antenna to a dual-antenna system. Keywords : Virtual multi-antenna array (VAA), RF Localization, Direction-of-arrival (DOA), Beamforming. | WG2 |
| TD(23)04042 | Bahram Khan, Nidhi, Rui R Paulo, Albena Mihovska, Fernando J. Velez | Cost Revenue Trade-off for the 5G NR Small Cell Network in the Sub-6 GHz Operating Band | 5G Radio Access Network (RAN) dis-aggregation has opened up opportunities toward the 2nd phase of 5G. 3GPP and Telecom industries have defined backhaul, fronthaul, and mid-haul transport interfaces, as well as functional splits to incorporate network flexibility and openness. In this work, splits 6 and 7 (7.2) of 3GPP are addressed for implementing sub-6 GHz future wireless mobile communication networks. The 5G-air-simulator has been considered to simulate New Radio 2.6 GHz, 3.5 GHz, and 5.62 GHz frequency bands by using Video (VI) and Video plus Best-Effort (VI+BE) with the Proportional Fair (PF) packet scheduler. The split 6 is ideal for small cell deployment, while split 7, (mainly sub-split 7.2) requires high fiber capacity, which may increase the price of the fronthaul. In the simulations, we have considered a uniform user distribution and reuse pattern three. By assuming a set of cost parameters and a given price for the traffic, we have analysed the cost/revenue trade-off of outdoor pico/micro cells, while comparing the implementation of functional splits 6 and 7 with scenarios without splitting. It is shown that, for all bands, for cell radii up to 500-600 m the split 6 and 7 provides higher revenue and profit compared to the case without splitting (with slight advantage for split 7). | WG3,VT4 |
| TD(23)04045 | Gianluca Fontanesi, Anding Zhu, Mahnaz Arvaneh, and Hamed Ahmadi, | A Transfer Learning Approach for UAV Path Design with Connectivity Outage Constraint | The connectivity-aware path design is crucial in the effective deployment of autonomous Unmanned Aerial Vehicles (UAVs). Recently, Reinforcement Learning (RL) algorithms have become the popular approach to solving this type of complex problem, but RL algorithms suffer slow convergence. In this paper, we propose a Transfer Learning (TL) approach, where we use a teacher policy previously trained in an old domain to boost the path learning of the agent in the new domain. As the exploration processes and the training continue, the agent refines the path design in the new domain based on the subsequent interactions with the environment. We evaluate our approach considering an old domain at sub-6 GHz and a new domain at millimeter Wave (mmWave). The teacher path policy, previously trained at sub-6 GHz path, is the solution to a connectivityaware path problem that we formulate as a constrained Markov Decision Process (CMDP). We employ a Lyapunov-based modelfree Deep Q-Network (DQN) to solve the path design at sub6 GHz that guarantees connectivity constraint satisfaction. We empirically demonstrate the effectiveness of our approach for different urban environment scenarios. The results demonstrate that our proposed approach is capable of reducing the training time considerably at mmWave. | WG3 |
| TD(23)04047 | SWARNA B. CHETTY, HAMED AHMADI , MASSIMO TORNATORE , AND AVISHEK NAG | Dynamic Decomposition of Service Function Chain Using a Deep Reinforcement Learning Approach | The Internet of Things (IoT) universe will continue to expand with the advent of the sixth generation of mobile networks (6G), which is expected to support applications and services with higher data rates, ultra-reliability, and lower latency compared to the fifth generation of mobile networks (5G). These new demanding 6G applications will introduce heavy load and strict performance requirements on the network. Network Function Virtualization (NFV) is a promising approach to handling these challenging requirements, but it also poses significant Resource Allocation (RA) challenges. Especially since 6G network services will be highly complicated and comparatively short-lived, network operators will be compelled to deploy these services in a flexible, on-demand, and agile manner. To address the aforementioned issues, microservice approaches are being investigated, in which the services are decomposed and loosely coupled, resulting in increased deployment flexibility and modularity. This study investigates a new RA approach for microservices-based NFV for efficient deployment and decomposition of Virtual Network Function (VNF) onto substrate networks. The decomposition of VNFs involves additional overheads, which have a detrimental impact on network resources; hence, finding the right balance of when and how much decomposition to allow is critical. Thus, we develop a criterion for determining the potential/candidate VNFs for decomposition and also the granularity of such decomposition. The joint problem of decomposition and efficient embedding of microservices is challenging to model and solve using exact mathematical models. Therefore, we implemented a Reinforcement Learning (RL) model using Double Deep Q-Learning, which revealed an almost 50% more normalized Service Acceptance Rate (SAR) for the microservice approach over the monolithic deployment of VNFs. | WG3 |
| TD(23)04049 | F. Munoz, L. Santamaria, T. Mazloum, A. Clemente, J.B. Gros, Y. Nasser, O. Mikhail, L. Geoffroy, R. D’Errico | Multiple RIS channel sounding: preliminary results on an experimental validation in RISE-6G project | In this TD we present the preliminary results of an experimental validation on the use of multiple RIS in indoor scenarios. The setup employs a Transmitting and a Reflecting RIS in the Ka band based on pin-diode technologies with more than 400 elements each. The results are analysed to prove the impact of RIS’ unit cells phase distributions on channel chracteristics. | WG1,Sub-WG1.2 |
| TD(23)04050 | Fred Wagen | About PHYsical Layer integration in system simulations: from channels models to abstract PHY and passthrough PHY | Accurate multi-user system level simulations and emulations appears to be increasingly important to optimize the use of radio spectrum resources. For example, scheduling the user services and devices depending on their requirements and propagation conditions could be optimized. Running multi-user system simulations might take several hours if all the details of the propagation conditions must be simulated. The MATLAB toolboxes, for both 5GNR and WiFi6 for example, offers system level simulations which simplify the effects of propagation impairments in various manner using namely (1) standardized multipath channel models, (2) abstract PHY: using PER-vs-SINR curve, and (3) passthrough PHY: e.g., using random CQI and a PER=0.9. This contribution presents an effort to understand the pros and cons of these approaches to analyze scheduling schemes. | WG3 |
| TD(23)04052 | Amélia Struyf, Cédric Hannotier and François Quitin | RF fingerprinting for wireless localization networks: comparison of Bayesian Multi-Target Tracking techniques | One main challenge of wireless localization networks is to recognise the nodes that transmit data packets, especially when the localization system does not decode MAC addresses. This paper presents a performance evaluation of different Bayesian Multi-Target Tracking (MTT) techniques for RF fingerprinting. The targets in this study transmit 802.11 Wi-Fi packets to a single receiver at regular time intervals. The Carrier Frequency Offset, considered as the fingerprinting feature, is extracted from these received packets. The system model accounts for the dynamics of the local oscillator (LO), which enables the inclusion of CFO drift in the analysis. This study compares and evaluates three different MTT methods: the Joint Probabilistic Data Association Filter (JPDAF), the Exact Nearest Neighbour version of the JPDAF (EN-NPDA) and Multi-Hypothesis Tracking (MHT). All of these methods rely on the use of Kalman filtering to track the state of multiple targets simultaneously. Different tracking scenarios are simulated to test the performance of the algorithms and highlight their limits. In addition, an experiment using Universal software-defined radios is performed to validate the algorithms. | WG2 |
| TD(23)04053 | M. Barbiroli, E. M. Vitucci, S. Kodra, V. Degli-Esposti, F. Fuschini, M. T. MartÍnez InglÉs, J. Molina Garcia Pardo, and S. Salous | Experimental Analysis of Mm-Wave Scattering and Attenuation from Construction Materials | Mm-wave frequency bands have been recently allocated to 5G communications and therefore the characterization of transmission, reflection and scattering characteristics of building materials at those frequencies is in the spotlight. Scattering from common construction materials is analysed in the paper through both measurements and simulations. | WG1 |
| TD(23)04056 | E. M. Vitucci, V. Degli-Esposti | Study of indoor RF Coverage solutions using Reconfigurable Intelligent Surfaces | A Huygens-based, antenna-array-like macroscopic model previously developed for scattering from metasurfaces is embedded into a ray tracing tool and used to perform RF coverage evaluations in a realistic indoor environment. Using the reciprocity of the link we extend prediction to multiple-bounce paths including RIS scattering. Different reference environments such as T-shaped and L-shaped corridor cases have been considered with different placement solutions of anomalous reflectors. Results show that a gain of about 10-15 dB can be obtained in blind spot locations with proper placement and configuration of the RIS. | Sub-WG1.2 |
| TD(23)04058 | Fernando José Velez, and Periklis Chatzimisios, Luis Orozco-Barbosa and Guillaume Villemaud | Challenges and Achievements in VT4 | This document provides a summary of the main identified challenges of VT4 by emphasising both on very high-rate communications and/or ultra-dense networks. Moreover, based on the TDs that have already been presented in the past two technical meetings, the document provides the first paths followed by the research activities of VT4. | VT4 |
| TD(23)04059 | Dino Pjanic, Alexandros Sopasakis, Andres Reial, Fredrik Tufvesson | Early-Triggered Handover Preparation in 5G NR Millimeter-Wave Systems | Parameters that directly influence radio signal transmission between the Radio Access Network and the User Equipment are optimized in layer 1 to 3. In these layers, algorithms perform millisecond-scale adaptations of the transmission parameters. In this paper we investigate how Machine Learning can be applied to time series data of beam measurements in Fifth Generation (5G) New Radio (NR) systems and improve procedures such as Handover. We show that beam measurements can be efficiently utilized as an early Handover trigger for inter-cell mobility optimization by predicting the remaining time to an imminent Handover event criteria fulfilment. Most current research studies focus on Handover optimization at a later point in time. As a proof-of-concept, we propose the novel, ML-enabled Early-Scheduled Handover Preparation scheme that offers user-context-aware Handover optimization. With these new insights, we also enable proactive, rather than reactive, decision-making during the Handover preparation phase in MIMO scenarios incorporating mobility. | WG3 |
| TD(23)04060 | Andre Ráth, Dino Pjanić, Bo Bernhardsson, Fredrik Tufvesson | ML-Enabled Outdoor User Positioning in 5G NR Systems via Uplink SRS Channel Estimates | Cellular user positioning is a promising service provided by Fifth Generation New Radio (5G NR) networks. Besides, Machine Learning (ML) techniques are foreseen to become an integrated part of 5G NR systems improving radio performance and reducing complexity. In this paper, we investigate ML techniques for positioning using 5G NR fingerprints consisting of uplink channel estimates from the physical layer channel. We show that it is possible to use Sounding Reference Signals (SRS) channel fingerprints to provide sufficient data to infer user position. Furthermore, we show that moderately Deep Neural Networks, even when applied to very sparse SRS data, can achieve successful outdoor user positioning with meter-level accuracy in a commercial 5G environment. | WG2,Sub-WG2 |
| TD(23)04061 | Roman Maršálek, Radim Zedka, Erich Zöchmann, Josef Vychodil, Radek Závorka, Ladislav Polák, Golsa Ghiaasi, Jiří Blumenstein | Persistent homology for human presence detection from 60 GHz Orthogonal Time Frequency Space transmissions | Orthogonal Time Frequency Space (OTFS) is a new modulation waveform, a promising candidate for integrated sensing and communication systems, providing the environment-awareness capabilities together with high speed wireless data communications. We present our 60 GHz measurement test-bed and results of OTFS-based person monitoring measurements in this millimeter-wave frequency band. We describe the use of the persistent homology technique as a method for processing of gathered delay-Doppler responses. The persistence homology approach is compared with the standard CFAR target detector for selected scenarios. | Sub-WG2,VT3 |