Dual connectivity is a key enabler for reliability and service continuity in 5G non-terrestrial networks, particularly in multi-band and multi-orbit satellite scenarios. However, its practical implementation at the user equipment side is challenged by implementation constraints. A primary obstacle is the timing advance mechanism: even with multiple radio interfaces, a software-defined radio-based UE can typically maintain only one valid timing advance value, making simultaneous bidirectional operation toward multiple cells infeasible. In this paper, we present a pragmatic UE architecture that overcomes this limitation to enable dual connectivity-like capabilities. Our design maintains one active link for bidirectional communication (both uplink and downlink), while a second radio interface is dedicated to establishing a monitoring link that maintains downlink-only synchronization and signal quality estimation of a secondary cell. Building on this fully UE-autonomous architecture, we implement a power level-based hard handover mechanism. The UE continuously monitors its active and monitoring links and triggers the handover based on local power level measurements, without any involvement from the base station or 5G core network. This approach provides a practical pathway toward improved link reliability in non-terrestrial systems while reducing network signaling using commercially available software-defined radio platforms.

The coexistence of terrestrial networks (TNs) and non-terrestrial networks (NTNs) within shared spectrum bands poses significant uplink (UL) interference challenges. While conventional studies assume single-antenna TN terminals, this work investigates UE-side precoding for NTN–TN coexistence using multi-antenna user equipment (UE) operating in upper mid-band frequencies, where larger antenna arrays at the UE become feasible. We develop a parametrized, direction-informed linear UL precoder that leverages antenna directivity and satellite ephemeris information, enabling UEs to shape their radiation patterns without requiring real-time satellite channel state information (CSI). By tuning two design parameters, the precoder can be configured for different practical objectives: an NTN-oriented setting that prioritizes interference suppression toward satellites, and a TN-oriented setting that preserves TN signal quality while adaptively prioritizing the most interference-exposed satellites, reducing the upper tail of the interference-to-noise ratio (INR) distribution and achieving an interference profile closer to uniform.

Managing radio resources reliably in networks that combine ground-based and satellite segments is an increasing challenge for next-generation mobile systems. AI-based traffic forecasting tools can predict future load, but they rarely indicate how much operators should trust a given prediction or what actions to take when different models disagree. This paper addresses that gap by introducing a lightweight trust layer that sits on top of existing forecasting models and combines their outputs with live network measurements to produce three simple indicators: how confident we are that a problem is coming (belief), how possible it is that things are still fine (plausibility), and how much the models disagree with each other (conflict). These three signals are then used to drive four practical actions: protecting capacity, maintaining the current state, hedging under uncertainty, and safely releasing unused resources, in a way that any network operator can inspect and adjust. Tested on real satellite traffic data, the approach eliminates service quality violations that occur with individual models, reduces unnecessary resource over-allocation from about 44% to just 5%, and lowers operational costs by roughly 87% compared to the best individual model alone. We also discuss how these ideas fit within current and future open radio network standards, and what open challenges remain before such a framework can be widely deployed.

This work investigates 2D distributed satellite swarm configurations for direct-to-cell (D2C) applications. Unlike previous studies, which primarily considered swarms as deployment alternatives to monolithic arrays, this paper focuses on exploiting the increased spatial resolution enabled by large distributed apertures. Simulation results show that, for the considered scenarios, increasing the level of antenna distribution across satellite platforms enlarges the effective aperture and improves the system sum rate under ideal operating conditions. While gains are moderate for uniformly distributed users, they become particularly pronounced in scenarios including hotspot with high user densities. The results further show that user scheduling strategies do not invalidate the superiority of highly distributed configurations, as these architectures provide a more balanced rate distribution across the coverage area, including hotspot regions. In addition, the paper analyzes several key implementation challenges associated with large distributed swarms, including errors in inter-satellite relative positioning, synchronization impairments, and limited beamforming and user-position update rates. Although within the range of swarm configurations and scenarios considered in this work, performance gains continue to increase with aperture size, these practical constraints may restrict swarm sizes in the medium term. Nevertheless, the observed performance gains strongly motivate further research into scalable synchronization, positioning and data distribution techniques for future large-scale satellite swarms.

The extension of 5G connectivity through Low-Earth Orbit satellite systems introduces significant technical challenges, particularly due to time-varying propagation delays and high Doppler shifts resulting from satellite motion. While the Third Generation Partnership Project Release 17 established the initial framework for non-terrestrial networks, the ongoing developments in Release 19 further enhance this effort by introducing support for regenerative payload architectures, where part of the communication protocol stack is processed directly on board the satellite. In this work, we present the design of a 5G user equipment adapted for Low-Earth Orbit satellite connectivity, with specific focus on strategies for managing variable delay and Doppler compensation. Additionally, we describe a custom experimental platform based on a drone-mounted software-defined radio platform capable of emulating both transparent and regenerative satellite payloads. Although full end-to-end system validation is not yet complete, initial laboratory tests confirm the feasibility of the architecture and lay the groundwork for future experimental campaigns.

Nonterrestrial networks (NTNs) are becoming a critical component of modern communication infrastructures, especially with the advent of low Earth orbit (LEO) satellite systems. Traditional centralized learning approaches face major challenges in such networks due to high latency, intermittent connectivity, and limited bandwidth. Federated learning (FL) is a promising alternative as it enables decentralized training while maintaining data privacy. However, existing FL models, such as FL with multilayer perceptrons (Fed-MLP), can struggle with high computational complexity and poor adaptability to dynamic NTN environments. This article provides a detailed analysis for FL with Kolmogorov-Arnold networks (Fed-KAN), its implementation, and performance improvements over traditional FL models in NTN environments for traffic forecasting. The proposed Fed-KAN is a novel approach that utilizes the functional approximation capabilities of KANs in an FL framework. We evaluate Fed-KAN compared to Fed-MLP on a traffic dataset of a real satellite operator and show a significant reduction in training and test loss. Our results show that Fed-KAN can achieve a 49% reduction in average test loss compared to Fed-MLP, highlighting its improved performance and better generalization ability. At the end of the article, we also discuss some potential applications of Fed-KAN within open radio access network (O-RAN) and Fed-KAN usage for split functionalities in NTN architecture.

This work evaluates the benefits of using large random single-antenna satellite swarms over traditional monolithically collocated single satellite arrays to serve handheld devices. It demonstrates that large random swarms can provide significant gains in interference limited regimes with large user densities, thanks to their reduced beamwidth and the moderate side lobes resulting from the randomized position of the satellite nodes. These findings have a direct application on the provision of satellite services to high user density hot spot areas. Besides, the operation with extreme narrow beamwidths opens new research challenges like efficient user scheduling or the full coverage of wide field of views.

Integrating non-terrestrial and terrestrial networks (NTN-TN) is essential to provide ubiquitous coverage in beyond-5G systems. This demo showcases a cloud-native testbed where an NTN UE and a TN UE communicate through a unified 5G core over an emulated geostationary satellite. The setup combines SDR hardware, realistic channel emulation, and automated lifecycle management of mobile network functions using ETSI NFV MANO and open-source platforms.

The integration of Terrestrial Networks (TNs) with Non-Terrestrial Networks (NTNs) poses unique architectural and functional challenges due to heterogeneous propagation conditions, dynamic topologies and limited onboard processing capabilities. This paper presents a taxonomy of architectural and functional split strategies for integrated Open Radio Access Network (O-RAN)-enabled TN-NTN systems. We analyze key trade-offs in performance, latency, and autonomy by evaluating configurations that distribute RAN and core functions between satellites and ground nodes, from onboard Distributed Unit (DU) deployments to full gNB and User Plane Function (UPF) integration. The placement of Near-Real Time (RT) and Non-RT RAN Intelligent Controllers (RICs) is also explored, with flexible strategies proposed to optimize control loop performance and scalability. We provide a comprehensive mapping between architectural splits and RIC placement options, highlighting implementation constraints and interoperability. Finally, we outline key challenges and future directions to facilitate efficient TN-NTN convergence within the O-RAN framework.

This paper presents a comprehensive performance evaluation of 5G Non-Terrestrial Networks (NTNs) as defined by the 3GPP Release 17 standard, with a focus on laboratory emulations and field trials using a geostationary satellite. We show initial laboratory tests, conducted with modems and a channel emulator, emulated various satellite orbits to assess communication latency, random access procedures, and system performance in controlled conditions. These tests provided valuable insights into potential communication challenges before transitioning to field trials. Field tests were performed with a dedicated transponder on a real geostationary satellite, evaluating performance metrics such as latency and throughput in both uplink and downlink channels. Results conducted with connection-less transport protocols indicate that while downlink throughput remains fairly stable across different altitudes, uplink performance degrades with increasing altitude, primarily due to increased latency from scheduling request delays. This trend is accentuated by employing connection-oriented transport protocols without accelerators. The findings confirm the feasibility of 5G NTN with satisfactory downlink performance, while highlighting the challenges in maintaining consistent uplink throughput at higher altitudes.

Efficient resource management is critical for Non-Terrestrial Networks (NTNs) to provide consistent, high-quality service in remote and under-served regions. While traditional single-point prediction methods, such as Long-Short Term Memory (LSTM), have been used in terrestrial networks, they often fall short in NTNs due to the complexity of satellite dynamics, signal latency and coverage variability. Probabilistic forecasting, which quantifies the uncertainties of the predictions, is a robust alternative. In this paper, we evaluate the application of probabilistic forecasting techniques, in particular Simple-Feed-Forward (SFF), to NTN resource allocation scenarios. Our results show their effectiveness in predicting bandwidth and capacity requirements in different NTN segments of probabilistic forecasting compared to single-point prediction techniques such as Long-Short Term Memory (LSTM). The results show the potential of probabilistic forecasting models to provide accurate and reliable predictions and to quantify their uncertainty, making them indispensable for optimizing NTN resource allocation. At the end of the paper, we also present application scenarios and a standardization roadmap for the use of probabilistic forecasting in integrated Terrestrial Network (TN)-NTN environments.

With the widespread adoption of 5G as a communication standard, satellite mega-constellations have emerged as viable alternatives and complement terrestrial networks, offering extensive and reliable communication services across a broad spectrum of users and applications. These constellations are already equipped with inter-satellite links and adaptable payloads capable of supporting Radio Access Network (RAN) and core network functionalities, forming complex space-based networks characterized by overlapping layers of multi-orbit, grid-like topologies that undergo continuous, yet predictable, changes—peculiarities not currently addressed within the 5G standards framework. To cope with this technology gap, this paper introduces a novel architecture for 5G services relying on satellite mega-constellations, which adhere to the principles of self-organized networks. This architecture is designed to align seamlessly with 5G service requirements, while also accommodating the unique topological and infrastructural constraints of mega-constellations. In more detail, the paper first outlines the fundamental principles of self-organizing networks that facilitate real-time system adaptation to internal topological shifts and external fluctuations in service demand. Then, we detail a 5G network architecture incorporating these principles, which includes 1) dynamic placement and migration of radio and core network control plane functions, 2) the strategic positioning of the data path, service, and AI decision functionalities to improve end-to-end service quality and reliability, and 3) the integration of dynamically established multi-connectivity options to increase the overall service dependability. These innovations aim for a seamless integration of space-based networks with terrestrial counterparts, creating a robust, cost-effective convergent telecommunication system.

This work proposes a pragmatic method for the design of beam footprint layouts and beam hopping illumination patterns to efficiently broadcast 3GPP NTN common signaling to large coverage areas using EIRP-limited LEO satellites. This method minimizes the time resources required to sweep over the whole coverage while ensuring that the signal-to-interference-plus-noise ratio received by users is above a given threshold. It discusses the design of: (i) an Earth-fixed grid of beam layouts; (ii) beamforming vectors and beam power allocation; (iii) beam hopping patterns and (iv) space, time and frequency resource allocation of 3GPP common signaling. Two main beam layout solutions are proposed to significantly reduce the number of beams required to illuminate the coverage area: one based on phased array beams with low beam crossover levels and the other on widened beams. A numerical evaluation using practical system parameters showed that both solutions perform similarly, but that the best result is obtained with phased arrays beams with optimized beam cross over levels. Indeed, for the system evaluated, they allowed reducing the total number of beams from 1723 to 451, which combined with a proper beam hopping pattern and scheduling scheme allowed obtaining a coverage ratio of 100

Low-Earth orbit (LEO) satellite networks, being a promising component of non-terrestrial networks (NTN), are meant to provide seamless and ubiquitous global connectivity. However, the high mobility of the LEO satellites results in frequent handovers (HOs), poses significant challenges to maintaining service continuity due to the interruption time experienced during the HO execution phase, which is particularly critical for high reliability communications (HRC). In this aspect, 3GPP proposed the dual active protocol stack (DAPS) HO mechanism for terrestrial networks (TN), a soft HO approach that enables the user equipment (UE) to maintain the simultaneous connection between both the source and target gNBs during the HO execution phase, potentially achieving a zero handover interruption time (HIT). However, DAPS HO is still not supported for NTN in latest 3GPP Release 18, likely due to rapid orbital movement of satellites. Moreover, DAPS HO requires synchronization between source and target nodes during the handover process, a condition that is challenging to identify at all times. Inspired by this challenge, this paper proposes a network-controlled algorithm that enhances the feasibility of DAPS HO in LEO satellites network, thereby minimizing the HIT over a given observation period. The proposed algorithm employs a sequential decision-making process within a predefined window to make the decision on configuring the DAPS HO. Simulation results demonstrate that the proposed approach significantly reduces the HIT, achieving a 80.28% reduction as compared to baseline HO scheme.

In this paper, we present an innovative federated learning (FL) approach that utilizes Kolmogorov-Arnold Networks (KANs) for classification tasks. By utilizing the adaptive activation capabilities of KANs in a federated framework, we aim to improve classification capabilities while preserving privacy. The study evaluates the performance of federated KANs (F-KANs) compared to traditional federated Multi-Layer Perceptrons (F-MLPs) on classification task. The results show that the F-KANs model significantly outperforms the F-MLP model in terms of accuracy, precision, recall, F1 score and stability, and achieves better performance, paving the way for more efficient and privacy-preserving predictive analytics.

The integration of semantic communication with Open Radio Access Network (O-RAN)-enabled Non-Terrestrial Networks (NTN) is a key enabler for 6G, optimizing bandwidth efficiency, Artificial Intelligence (AI)-native inference and intelligent satellite-terrestrial integration. Unlike traditional bit-level transmission, semantic-aware networks extract and transmit only meaningful information, reducing overhead and improving the adaptability. This paper presents SEM-NTN, a semantic-aware O-RAN-enabled NTN framework, that leverages AI-driven feature extraction, adaptive compression, and dynamic resource allocation. Simulation results show that SEM-NTN reduces transmission delay by up to 87.5% while maintaining AI inference accuracy close to that of full-quality compression. Notably, semantic-aware compression achieves up to 84.6% mean Average Precision (mAP) and 77.6% mean Intersection over Union (mIoU) under a 5ms delay constraint-closely matching the uniform baseline (85% mAP, 78% mIoU) but with significantly improved latency performance. These findings highlight SEM-NTN's potential for scalable, low-latency, and resource-efficient 6G communications. The paper concludes by outlining key challenges in semantic protocol design, real-time adaptation, and standardization for AI-native network integration.

The increasing spectrum overlap between non-terrestrial networks (NTNs) and terrestrial networks (TNs) introduces significant uplink (UL) interference challenges. This paper addresses NTN–TN coexistence by exploiting the spatial degrees of freedom available at TN user equipment (UE) with large antenna arrays, which become feasible at upper mid-band frequencies. We propose an uplink precoding method based on regularized zero-forcing (RZF), adapted to suppress interference toward satellites while preserving TN performance. Unlike classical RZF applications at the base station, the proposed scheme is applied at the UE and does not require real-time estimation of satellite channel state information (CSI). Moreover, the gNB centrally computes a user-specific regularization parameter. To address hardware constraints, hybrid beamforming is also incorporated. The proposed approach is benchmarked against Maximum Ratio Transmission (MRT) and the interference nulling scheme. Simulation results show that adaptive RZF significantly reduces interference at satellites while preserving terrestrial signal quality, offering an effective solution for NTN–TN spectrum sharing in the upper mid-band.

This work introduces Probabilistic Kolmogorov-Arnold Network (P-KAN), a novel probabilistic extension of Kolmogorov-Arnold Networks (KANs) for time series forecasting. By replacing scalar weights with spline-based functional connections and directly parameterizing predictive distributions, P-KANs offer expressive yet parameter-efficient models capable of capturing nonlinear and heavy-tailed dynamics. We evaluate P-KANs on satellite traffic forecasting, where uncertainty-aware predictions enable dynamic thresholding for resource allocation. Results show that P-KANs consistently outperform Multi Layer Perceptron (MLP) baselines in both accuracy and calibration, achieving superior efficiency-risk trade-offs while using significantly fewer parameters. We build up P-KANs on two distributions, namely Gaussian and Student-t distributions. The Gaussian variant provides robust, conservative forecasts suitable for safety-critical scenarios, whereas the Student-t variant yields sharper distributions that improve efficiency under stable demand. These findings establish P-KANs as a powerful framework for probabilistic forecasting with direct applicability to satellite communications and other resource-constrained domains.

Future 6G communications are expected to be complemented by non-terrestrial networks (NTNs), particularly satellites, to ensure ubiquitous connectivity. Furthermore, to address the growing spectrum shortage in terrestrial networks (TNs), the upper mid-band (7-24 GHz), currently known as frequency range 3 (FR3), is emerging as a key frequency band due to its appealing balance between capacity and coverage. However, this band is already utilized by incumbent satellite communications, making the investigation of NTN and TN coexistence in FR3 critical. This paper explores linear multi-antenna receiver techniques at TN user equipment (UE), aimed at enhancing coexistence between NTN and TN, leveraging the increased number of antennas that higher frequency bands enable on UE. Specifically, we apply Maximum Ratio Combining (MRC) and Interference Rejection Combining (IRC) techniques to address the dual challenges of interference suppression and signal enhancement. Simulation results demonstrate the effectiveness of these techniques in improving TN UE performance and reducing satellite interference.