Wireless connectivity underpins modern society and industry, enabling critical applications such as 5G ultra-reliable low-latency communication (URLLC) for industrial automation. However, the openness of the wireless medium exposes it to spectrum anomalies, including unintentional interference and malicious jamming, which threaten communication and sensing functionalities in 5G and emerging 6G networks. Despite its importance, spectrum anomaly detection research is hindered by a lack of publicly available datasets reflecting real-world scenarios. To address this, we present a benchmark dataset for spectrum anomaly detection in orthogonal frequency-division multiplexing access (OFDMA) systems, a core technology for 5G and beyond. The dataset includes spectrograms generated across a distributed network of sensing units, covering five distinct jammer types, from simple noise to advanced pilot-aware attacks. These anomalies are simulated in an industrial factory environment using a versatile open-source framework developed and published as part of this work, enabling extensibility to new scenarios and interference types. We provide baseline evaluations for supervised and unsupervised learning methods, demonstrating the challenges posed by different jammers and highlighting areas for further research. The dataset and framework support reproducible studies and serve as a foundation for advancing spectrum anomaly detection, with applications extending to network digital twins. By bridging the gap in open dataset availability, this work empowers the research community to validate and compare advanced detection methods for resilient next-generation wireless systems.

This paper highlights key areas in which AI/ML can play a transformative role in 6G, with an emphasis on energy-efficient solutions. Contextualized within Germany’s national 6G research initiative, "6G Access, Network of Networks, Automation and Simplification" (6G-ANNA) emerges as the lighthouse project, providing a holistic vision that sets the direction for numerous specialized research efforts. Within this context, this paper aims to present ideas relevant to both academia and industry by identifying key research directions for AI/ML. We begin by reviewing the latest developments in using AI/ML for the Fifth Generation (5G) New Radio (NR) air interface as discussed in 3rd Generation Partnership Project (3GPP) Releases 18 and 19, and examine how these advancements pave the way for a native and energy-efficient AI/ML air interface in 6G. Key results are presented on AI/ML-driven optimization of radio frequency (RF) frontends, along with a strong focus on the role of AI/ML in diverse signal processing tasks and energy saving mechanisms, which demonstrate the potential of AI/ML in improving spectral efficiency and reducing energy consumption. The discussion further introduces methodologies for testing AI/ML-based signal processing tailored for the 6G physical layer, addressing practical challenges relevant to industry stakeholders and standard development organizations. Finally, we discuss the standardization aspects critical for realizing a future AI-native air interface in 6G, aligning our findings with ongoing and upcoming global standardization activities.

The integration of extremely large (XL) antenna arrays with ultra wideband (UWB) communication is pivotal for next-generation networks. However, the multitude of candidate beams renders conventional analog phase shifters (APSs)-based training prohibitively costly, necessitating true-time delay (TTD)-based frequency-dependent beamforming. We design a codebook mapping distinct orthogonal frequency-division multiplexing (OFDM) subcarriers to unique spatial angles, enabling full angular coverage within a single OFDM symbol. To address practical TTD limitations-finite delay range and resolution-we propose a multi-stage delay-phase-precoding (MSDPP) architecture that mitigates hardware-induced distortions. Simulations demonstrate that MSDPP achieves near-optimal spectral efficiency (SE) with at least a 2dB gain over state-of-the-art designs, alongside up to 63% higher energy efficiency (EE).

Ensuring high throughput and stable connectivity for mobile user equipment (UEs) with multiple receiver antenna panels is a key challenge in modern mobile networks. It is well-established that current handover mechanisms, based on procedures from the Third Generation Partnership Project (3GPP), exhibit performance limitations, especially in scenarios with high mobility at cell edges or at large distances from the base station (BS). In these areas, the strong influence of shadowing is known to cause an excessive number of unnecessary handovers (HOs), which significantly reduce throughput and reliability. To overcome this limitation, we introduce a novel framework based on Deep Q-Networks (DQN) to optimize handover decisions in Non-Coherent Joint Transmission (NCJT) mobile scenarios. Our approach enables the UE to control the selection of the transmission reception points (TRPs) autonomously and intelligently. We design and train the DQN for two different operating modes: one designed for maximum throughput and one for maximum reliability. The simulation results confirm that the DQN-based approach increases the overall performance compared to the 3GPP legacy procedures in critical scenarios by up to 60 %. These results establish deep reinforcement learning as a powerful tool for adaptive mobility management of multi-panel UEs and open up promising application possibilities for future wireless systems that require intelligent and context-sensitive TRP selection.

Extremely large-scale (XL) antenna arrays and wideband transmission are critical enablers for next-generation mobile communication networks. These technologies offer substantial improvements in angular resolution, spatial degrees of freedom, and spectral efficiency (SE). However, the combination of large antenna apertures and ultra-wide bandwidth induces a frequency-dependent beam misalignment, known as the beam squint effect, which severely degrades spatial directivity and SE. Traditional frequency-independent analog phase shifters (APSs) cannot compensate for this phenomenon, necessitating the use of true-time delays (TTDs). Despite their potential, most existing TTD-based architectures rely on the assumption of ideal hardware, ignoring practical constraints such as limited delay range and finite delay resolution. This paper analyzes the performance of existing architectures under these practical hardware limitations and proposes a novel hardware-aware multi-stage delay-phase-precoding (MSDPP) architecture. The proposed MSDPP is designed to comply with commercially available device specifications, effectively minimizing the impact of quantization and clipping errors inherent in practical TTDs. Extensive simulations considering realistic hardware constraints demonstrate that the proposed MSDPP outperforms state-of-the-art solutions by approximately 3 dB and 55 % in SE and energy efficiency (EE), respectively, making it a robust solution for energy-efficient ultra-wide bandwidth extremely large-scale XL antenna array systems.

The evolution towards sixth-generation (6G) and vehicle-to-everything (V2X) networks requires precise quality-of-service (QoS) levels, making network behavior prediction crucial for applications like autonomous driving. While machine learning (ML) is a promising solution, its performance is highly dependent on hyperparameter tuning, a complex and computationally expensive challenge. This study proposes a rigorous methodology for tuning Random Forest (RF) hyperparameters for downlink throughput prediction in vehicular communication scenarios using the Berlin vehicle-to-everything (V2X) dataset, collected in West Berlin (Germany). The core of the methodology is a novel statistical approach named Analysis of Variance and Scott-Knott (ANOVASK), which utilizes Analysis of Variance (ANOVA) and the Scott-Knott clustering algorithm to systematically rank combinations of eight RF hyperparameters. Two distinct ANOVA models: a one-way and a three-way factorial design were employed to assess the influence of hyperparameter values on Artificial Intelligence (AI) model performance. Results revealed that different hyperparameter settings lead to statistically significant variations in prediction accuracy. When compared with traditional methods, the models recommended by the ANOVASK algorithm outperformed both Random Search and Grid Search , with discrepancies not exceeding 2.2% in Mean Absolute Error (MAE), 5.1% in Root Mean Squared Error (RMSE), and 0.8% in Coefficient of Determination (R2). The proposed approach also demonstrated superior computational efficiency, requiring 14 to 16.4 h to obtain the models, compared to the 24 h taken by the other methods. Overall, the proposed methodology provides a robust, accurate, and efficient solution for RF hyperparameter tuning in vehicular networks.

As we increasingly rely on wireless infrastructure for enabling critical applications, e.g., from Industrial Internet of Things (IIoT) or healthcare, we must ensure resilience, which includes robustness against both accidental and intentional jamming attacks. Hence, awareness of the potential jamming attack information is crucial to the resilience of a fully automated system. However, channel estimation within modern communications remains highly vulnerable to such attacks. The jamming signal is considered to be projected onto the captured channels, which enables the jamming detection through channel analysis. In this work, we consider the detection of a jamming source within an indoor environment. Narrowband white Gaussian noise (WGN) is employed to jam the communications. The impact of the jamming signal on the channel measurements is analyzed by examining the correlation between the jammed and unjammed channels. Based on the correlation pattern, the presence of interference is detected through a comparative suite of statistical thresholding and machine learning algorithms. An overall detection accuracy of 99.31% is achieved, demonstrating the feasibility of jamming detection through the proposed framework.

Ultra-reliable low-latency communications (URLLC) demands resource allocation strategies that balance reliability with computational efficiency. Existing theoretically optimal methods suffer from prohibitive complexity for real-time deployments, while heuristic alternatives remain computationally demanding for large-scale systems. This paper proposes a swapping-based framework that exploits temporal channel correlation and frequency selectivity to distribute allocation overhead across time and frequency domains. The approach selects a limited subset of bottleneck resources, establishes their exchange priorities based on channel quality metrics, and performs selective swaps to enhance user performance. Our evaluation using real-world industrial channel measurements demonstrates that the proposed method achieves comparable reliability to legacy approaches while reducing allocation time by orders of magnitude, with computational complexity that scales gracefully across varying system scales.

Future 6G networks must manage increasingly dynamic radio environments in which multiple autonomous sub-networks (SNs) share frequency resources and adapt to changing operating conditions. In such scenarios, interference from faulty devices or intentional jamming can disrupt ongoing communications, making rapid and autonomous network adaptation essential. This demonstration presents a self-healing networks-in-network (NiN) architecture that closely integrates the detection of spectrum anomalies with dynamic spectrum management. A spectrum scanner continuously monitors the frequency spectrum and forwards detected anomalies to the DSM, which automatically identifies suitable frequency resources and reconfigures the affected SN. During the live demonstration, participants can initiate controlled disruptions and observe the entire adaptation process in real time, from anomaly detection to autonomous frequency reallocation and network recovery. The demonstrator illustrates how integrating spectrum monitoring and resource management into a single control loop can improve the resilience of future NiN implementations and demonstrates a practical approach to autonomous, spectrum-aware networking.

Ultra-reliable low-latency communications (URLLC) over Wi-Fi orthogonal frequency-division multiple access (OFDMA) networks is challenged by rapid channel fluctuations and the high complexity of conventional Max-Min resource allocation. This paper proposes an adaptive Max-Min framework that continuously monitors channel quality, triggering optimization only when reliability degrades and restricting each update to a carefully selected subset of users. This demand-driven strategy preserves Max-Min fairness while reducing computational effort and signaling overhead by an order of magnitude compared to traditional periodic Max-Min. It also outperforms representative heuristic allocators. Validation using measured industrial channel traces demonstrates that our approach sustains URLLC-grade reliability and latency under realistic mobility and spectrum conditions.

-6th generation cellular networks are foreseen to enable various mission-critical use cases, not only by means of communications, but also by the integration of sensing capabilities, referred to as joint communications and sensing. However, both communications and sensing are sensitive to interfering signals that can degrade communication or sensing performance. To ensure spectrum integrity in non-public networks (NPNs), we propose leveraging a radio environment digital twin (DT) to compare the channel frequency response (CFR) of the real-world wireless channel with that of the digital twin, referred to as digital twin channel (DTC). Significant deviations between the two may indicate spectrum anomalies, i.e., any kind of interfering signal. In an evaluation on measurement data collected from a workshop environment, the proposed approach achieves a missed detection rate below 0.1% for a tolerated false positive rate (FPR) of 1%. Moreover, non-smooth regions in the CFR can serve as an additional indicator of spectrum anomalies. Using the magnitude of the forward difference as a measure of non-smoothness and combine the two approaches, the missed detection rate can be further reduced to $4 \cdot 10^{-5}$ while maintaining a constant FPR. Thus, this paper demonstrates that integrating contextual awareness, enabled by DT, facilitates high-fidelity spectrum anomaly detection, a critical capability for building resilient future NPN.

Future industrial shopfloors will feature a large number of autonomous, mobile devices that communicate wirelessly. In such controlled environments, collecting channel state information (CSI) by location and reusing it over time may reduce the overhead of channel sounding and CSI feedback that is required for channel-aware radio resource allocation. However, even if the trajectories of devices repeat over time, temporary variations of the environment put a question mark behind the usability of radio environment maps (REMs) from the past. To address this important question, we designed and carried out a channel measurement campaign in an industrial-like area, we evaluated the data regarding the consistency of the radio channel in a time-varying environment, and we conducted simulations of a multi-user communications scenario based on measured data to present a potential application of channel information reuse. The measurement results show a high similarity of the REM over time and a spatially limited impact of the temporary variations caused by a metallic object on the REM. Our simulation results show up to 75% savings in CSI feedback overhead while the reliability of the communications system remains almost unaffected. In the future, we will elaborate on the estimation of the REM in a time-varying environment.

The sixth generation (6G) of mobile networks is being developed to overcome limitations in previous generations and meet emerging user demands. As spearhead of the European research and development effort on 6G, the Smart Networks and Services Joint Undertaking (SNS JU) 6G Flagship project Hexa-X-II has a leading role for developing the technologies and anchoring 6G end-toend (E2E) system. This paper summarizes the security, privacy, and resilience (SPR) controls identified by the Hexa-X-II project and their validation activities. Moreover, we share the SPR view on the 6G E2E system with the SPR features which are necessary to ensure the trustworthiness of 6 G.

The rapid evolution of cellular networks, driven by the proliferation of mobile devices and the exponential growth of the Internet of Things (IoT), has significantly advanced wireless communication technologies. Fifth generation of wireless communications technology (5G) enhanced data rates, latency, and network capacity, resulting in the emergence of new applications. However, the sixth generation (6G) is foreseen to support a new set of use cases with diverse requirements. This paper explores the critical role of artificial intelligence (AI) in shaping the trajectory from 5G to 6G. We discuss AI applications in 5G for network planning, resource allocation, traffic management, and security, as well as propose infrastructure upgrades, like edge servers and enhanced network topologies, to support AI in 6G. Additionally, we outline a visionary perspective on AI’s potential contributions to 6G, highlighting its role in enabling innovative services and applications. By providing this forward-looking perspective, this paper aims to stimulate discussion and guide the development of intelligent and autonomous 6G networks.

Reconfigurable intelligent surfaces (RISs) are emerging as a transformative technology capable of shaping the wireless communications channel for favorable propagation. By passively beamforming wireless signals in the desired directions, RISs hold great potential to enhance network coverage and reliability of wireless channels. Nevertheless, several unaddressed challenges hinder its practical deployment in the cellular context. This paper identifies six pivotal challenges that must be resolved to ensure the successful integration of RISs into wireless communication networks. Further, we analyze the influence of different frequency regimes on the severity of these challenges. Finally, we highlight conditions under which RISs can deliver substantial benefits and stand out as a key enabler for next-generation wireless communications networks.

The massive availability of wireless links will shape the factory of the future. Many envisioned wireless use cases rely on ultra-reliable low latency communication (URLLC). While providing single devices with URLLC is trivial these days, scaling URLLC networks up to many users remains a challenge. Channel-aware radio resource allocation is a method to face this challenge efficiently. However, fine-granular channel awareness is required for channel-aware radio resource allocation. A recently proposed approach to address this requirement is to reuse the channel state information (CSI) stored in a radio environment map (REM). To successfully apply the information from the REM, very accurate localization of the radio device(s) is required, which is the focus of this paper. Based on measurement data and spatially extremely fine-grained fingerprint sequences, it is demonstrated that localizing an automated guided vehicle (AGV) as a radio device is possible with sub-centimeter accuracy on its trajectory. The performance is achieved with the k-nearest neighbor (kNN) approach, and its robustness, even in a time-varying environment, is highlighted. As an enabler for channel-aware resource allocation, the presented work represents a step towards scaling up URLLC networks in industrial use cases.

Rigorous and reliable alignment of narrow transceiver beams is a requisite for ensuring the highly directional transmission in millimeter-wave (mmWave) communications. Exhaustively testing these narrow beam pairs results in increased reference signal (RS) overhead, latency, and power consumption. In this paper, we propose a centralized multi-task learning (MTL) based beam prediction strategy that ensures a high success rate using measurements from a few site-specific probing beams identified via the proposed uniformly distributed beam relevance and beam significance (UDBRBS) criterion, thereby obviating the need for an exhaustive scan. Performance evaluation over 3rd Generation Partnership Project (3GPP) defined performance indicators demonstrates that the proposed method outperforms existing independent task learning (ITL) and single task learning (STL) beam prediction designs. We further argue that the proposed strategy is highly practical for implementation in fifth generation (5G)-Advanced and sixth generation (6G) communication systems

With the increasing adoption of orthogonal frequency division multiple access (OFDMA) technology in WiFi networks, achieving ultra-reliable low-latency communications (URLLC) presents new challenges in resource management. We propose implicit channel quality-driven resource allocation (ICQRA), a lightweight framework that avoids explicit channel sounding overhead while maintaining high reliability. ICQRA passively monitors channel quality during normal packet demodulation, proactively detects degradations, and switches to better frequency resources before packet losses occur. Our approach incorporates an adaptive threshold mechanism that learns from transmission outcomes while efficiently coordinating resources across multiple users. Evaluations using both simulations and industrial channel measurements demonstrate that ICQRA achieves near-optimal performance in reducing consecutive packet losses with significantly lower resource consumption compared to existing approaches.

Large antenna apertures and/or ultra wideband (UWB) transmission engender the spatial wideband (SWB) effect, resulting in reduced array gain, spectral efficiency (SE), and higher bit error probability. To this end, this letter proposes a novel double-delay-phase precoding (DDPP) architecture that can be realized with a small number of true time delays (TTDs), mitigating the negative impact of the SWB effect. The numerical results, over the massive and extreme multiple-input multiple-output (MIMO) use cases and in the frequency range (FR)2 and FR3, showcase the effectiveness of the proposed DDPP in terms of the array gain, SE, and bit error rate (BER).

With the increasing adoption of orthogonal frequency division multiple access (OFDMA) technology in WiFi networks, accurately modeling packet loss is crucial for enhancing communication reliability. This is particularly important for ultra-reliable low-latency communications (URLLC), where the complexity of channel models can be prohibitive with many realizations. This paper introduces two streamlined approaches: a correlation-based model and a multi-dimensional Markov chain model. Both models are engineered to capture the intricate interplay of time and frequency dependencies among resource units in OFDMA WiFi systems. By leveraging frequency diversity and employing selection combining techniques, this study validates the accuracy of the proposed models in analyzing packet loss. The paper presents efficient model approximations that establish a comprehensive framework to simulate and analyze packet loss dynamics, contributing to the enhancement of network reliability and the optimization of system performance.