6G-SHINE: ADVANCING CORE TECHNOLOGIES FOR IN-X SUBNETWORKS

WHITE PAPER 6G-SHINE: ADVANCING CORE TECHNOLOGIES FOR IN-X SUBNETWORKS OCTOBER 2025 DOI: 10.5281/zenodo.17338153 URL: https://doi.org/10.5281/zenodo.17338153 2 TABLE OF CONTENTS EXECUTIVE SUMMARY ............................................................................... 3 ACKNOWLEDGEMENT .................................................................................................. 3 ABBREVIATIONS ......................................................................................... 4 1. INTRODUCTION ...................................................................................... 5 2. USE CASES FOR IN-X SUBNETWORKS ...................................................... 5 2.1 ARCHITECTURAL ENABLERS ................................................................................... 9 3. IN-X SHORT-RANGE RADIO CHANNEL CHARACTERIZATION .................. 12 4. PHYSICAL LAYER AND MEDIUM ACCESS CONTROL ENABLERS .............. 13 5. RADIO RESOURCE MANAGEMENT ........................................................ 16 6. MANAGEMENT OF TRAFFIC, SPECTRUM AND COMPUTATIONAL RESOURCES .............................................................................................. 18 7. DEVELOPED PROOF-OF-CONCEPTS ....................................................... 19 8. OUTLOOK ............................................................................................. 20 3 EXECUTIVE SUMMARY The 6G-SHINE project set out to design the foundational technology components that will enable the next generation of in-X subnetworks—short-range, low-power radio cells operating at the very edge of the 6G “network of networks.” These subnetworks are envisioned to provide pervasive wireless coverage with unprecedented performance requirements. They should be able to operate independently, much like ad hoc networks, with the advantages of seamless integration into the broader 6G system, which can support spectrum and computational resource management, authentication and authorization, traffic policy enforcement, and data offloading. The project advanced research across multiple technical domains, including the physical layer, medium access control, radio resource management, and network architecture. The ambition was to create a portfolio of innovations that can transform in-X subnetworks from concept into reality. This involved not only developing mechanisms for local communication but also establishing how such subnetworks can be efficiently managed, secured, and interconnected with the wider 6G framework. Through this work, 6G-SHINE has laid the groundwork for future wireless systems in which groups of devices can form intelligent, autonomous, and highly efficient subnetworks operations. These advances open the door to new applications in consumer, industrial, and vehicular domains, where local connectivity, robustness, and responsiveness are critical. This whitepaper is based on the 6G-SHINE project deliverables and presents a concise overview of the main innovation developed in the project. Acknowledgement This whitepaper is based on the deliverables of the 6G-SHINE project (Horizon Europe Grant Agreement No. 101095738), publicly available in the project website. This project is funded by the European Union. Views and opinions expressed here are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Health and Digital Executive Agency. Neither the European Union nor granting agency can be held responsible for them. 4 ABBREVIATIONS Abbreviation Meaning 3GPP 3rd Generation Partnership Project AI Artificial Intelligence AR Augmented reality BLER Block Error Rate CN Core Network CSI Channel State Information DNN Deep Neural Networks ECU Electronic Control Unit GDRA Gradient Descent-based Resource Allocation GNN Graph Neural Networks GW Gateway Role HC Element with High Capabilities HPCU High Processing Control Unit ISAC Integrated Sensing and Communication LC Element with Low Capabilities LOS Line-of-Sight LSTM Long Short-Term Memory MAC Medium Access Control ML Machine Learning NLOS Non Line-of-Sight O-RAN Open Radio Access Network OFF Compute Offloading PoC Proof-of-Concept PPO Proximal Policy Optimization RAN Radio Access Network RIS Reconfigurable Intelligent Surface RRM Radio Resource Management SL Sidelink SN In-X Wireless Subnetwork SNE Subnetwork Element SNM Subnetwork Management SNR Signal-to-noise ratio TRL Technology Readiness Level UE User Equipment URLLC Ultra-reliable Low Latency Communication VAA Vehicle Access Area XR Extended Reality 5 1. INTRODUCTION The 6G Short range extreme communication IN Entities (6G-SHINE) project, funded under the SNS Work Programme Phase 1 Stream B, targets the design of core technology components for wireless in-X subnetworks—short-range, low-power radio cells deployed at the very edge of the 6G “network of networks.” These subnetworks aim to deliver pervasive wireless coverage while meeting unprecedented performance requirements. They should be able to operate independently similarly to ad hoc networks, while benefitting from connection with the 6G parent network, that can ease management of spectrum and computational resources, authentication / authorization, policy enforcement for traffic to/from devices in the subnetwork, traffic offloading. 6G-SHINE advanced research in the physical layer, medium access control, radio resource management, and network architecture, with the ambition of assembling a portfolio of technology components that can turn in-X subnetworks from concept into reality. The project brought together a consortium of 12 partners: Aalborg University (Denmark), CNIT (Italy), Universidad Miguel Hernández de Elche (Spain), Apple (Germany), Sony (Sweden), Bosch (Germany), Interdigital (UK), Fraunhofer IIS (Germany), imec (Belgium), Keysight (Finland), Nokia (Denmark), and Cogninn (Greece). In this whitepaper, we present a concise overview of the innovative ideas and concepts developed in the project. Section 2 presents the main use cases envisioned for in-X subnetworks, covering the consumer, industrial, and in-vehicle domains, along with the main architectural enablers. Section 3 summarizes the radio channel characterization activities run in the project, covering diverse use cases and frequency bands. The developed physical layer and medium access control enablers are presented in section 4, while section 5 focuses on radio resource management research. Section 6 presents our research outcome for the management of traffic, spectrum and computational resources. The realized PoCs are presented in section 7. Finally, Section 8 provides concluding remarks, and an outlook for future activities. All project deliverables are available on the project website (6gshine.eu), with references provided throughout this whitepaper to link the innovations presented to their corresponding deliverables. For a comprehensive description of the innovations, the reader can refer to the project deliverables. 2. USE CASES FOR IN-X SUBNETWORKS In-X subnetworks are short-range, low power radio cells to be installed in entities such as robots, production modules, vehicles, and classrooms, to deliver highly localized and high-performance connectivity. We have identified 13 relevant use cases for in-X subnetworks, covering the consumer, industrial and in-vehicle domains. The use cases, pictorially depicted in Figure 1, 2 and 3, have extensively been described in deliverable D2.2, along with their performance requirements, and are summarized in Table 1. 6 Table 1 6G-SHINE in-X subnetworks use cases No Title Short Descriptions C -1 Immersive Education Immersive Education aims to enhance the interactive experience for a group of students and teacher(s) for knowledge exchange, leveraging media content and related technologies (e.g., XR devices). C -2 Indoor Interactive Games XR interactive gaming in an indoor environment where one or more players play in a place where it has been equipped and pre - loaded with some equipment to facilitate the XR interactive gaming. C -3 Virtual Live Production One or more performers that can be located in different geographical areas produce 3D video content that can be live - broadcasted or uploaded to social media. C -4 AR Navigation AR navigation powered by AI/ML concierge based on user input, including the information around the user. The output is provided to the user via an AR device. I -1 Robot Control The wireless control of robot operations, such as the control of multi - axis robots for leveraging the degrees of freedom offered by potential movement directions the robot can accomplish. I -2 Unit Test Cell To perform quality assurance tasks of product parts in the manufacturing process, as well as of devices used in the manufacturing process. I -3 Visual Inspection Cell A visual inspection cell performs quality assurance in the manufacturing process through video feeds. The video feeds are processed, and quality control is performed, by eventually outputting commands to actuators in case actions are to be taken for improv ing operation quality. I -4 Subnet Co - existence in Factory Hall Tasks distributions among a swarm of smaller, specialized robots. Each robot is configured to perform a specific function. Working in concert, these robotic swarms can assemble intricate products I -5 Subnetwork Segmentation and Management Combination of all the other industrial use cases with a focus on the security and management aspects, particularly on functional requirements. 7 V -1 Wireless Zone ECU: in -vehicle wireless subnetwork zone In - vehicle zone wireless subnetwork that is utilized by some sensors and actuators located in this zone to connect wirelessly to the zone Electronic Control Unit (ECU) that manages and controls them. The sensors and actuators that are wirelessly connected to the zone ECU are equipped with a 6G - capable wireless communication interface that replaces their former wired communication interface. V -2 Collaborative Wireless Zone ECUs: Functions across multiple in -vehicle zones This use case covers automotive systems and applications that require (or benefit from) collaboration or offloading between functions, sensors and actuators located at different zones of the considered 6G -SHINE reference invehicle E/E architecture , via a High P rocessing Control Unit (HP CU). V -3 Inter -subnetwork Coordination: Collaboration/interference/RRM between subnetworks in intra/inter -vehicle communications Two cases of subnetwork coordination on RRM/interference handling where the subnetworks can be within the same vehicle (Intravehicular) or in some other vehicle(s) (Intervehicular). V -4 Virtual ECU: In -vehicle sensor data and functions processing at the 6G network edge Integrating the in - vehicle network with the 6G parent network, to seamlessly extend the invehicle embedded compute capabilities to the edge/cloud. Figure 1. Use cases for consumer subnetworks. 8 Figure 2. Use cases for industrial subnetworks. Figure 3. Use cases for in-vehicle subnetworks. Consumer Subnetworks: Use cases in consumer subnetworks typically demand high data rates and low latency. Computation offloading is commonly employed, particularly for video rendering and AI/ML data processing. A baseline level of reliability is essential to minimize retransmissions, which could otherwise increase latency. These scenarios involve a diverse set of KPIs across different links—such as one or more high-throughput data streams (e.g., uplink/downlink video frames) accompanied by multiple low-rate sensor data streams. Additionally, synchronization among multiple media streams—including video, audio, and sensor data—is often required. Meeting the simultaneous demands of high throughput (up to 6.37 Gbps), low latency, and high reliability presents significant technical challenges. 9 Industrial Subnetworks: Industrial use cases generally prioritize ultra-low latency and high reliability. Certain scenarios require communication cycles below 100 µs. Although the data rates per device are typically low, the number of connected devices per access point (AP) can be high. Furthermore, multiple APs may need to operate concurrently, necessitating a welldesigned network architecture to support stringent latency requirements and enable potential computation offloading between nodes. It is worth noting that industrial use cases involving video feeds may also require relatively high data rates—up to approximately 50 Mbps. In-Vehicle Subnetworks: In-vehicle use cases place a strong emphasis on extreme reliability, with some scenarios demanding up to 99.9% reliability. These deployments often involve a mix of mediumrate links—such as those carrying image or video data from cameras—and low-rate links related to interactions with various sensors and actuators. The use cases defined within the 6G-SHINE framework are expected to contribute meaningfully to key value indicators (KVIs) related to social, environmental, and economic sustainability. Social sustainability is addressed through enhancements in user experience, ensuring that future connectivity services deliver greater inclusivity, accessibility, and quality of life compared to existing technologies. Environmental sustainability is pursued through targeted reductions in carbon emissions, including both carbon dioxide and methane, alongside a decrease in overall energy consumption across network operations and end-user applications. Economic sustainability is addressed through different mechanisms depending on the subnetwork context. In industrial subnetworks, the focus is on improving production efficiency, thereby enhancing competitiveness and resource utilization. In in-vehicle subnetworks, economic benefits stem from the extension of vehicle lifecycles and the promotion of circular economy principles. In consumer subnetworks, economic sustainability is supported by enabling new digital services in sectors such as education and gaming, potentially unlocking new markets and innovation opportunities. Across all use cases, low-latency and reliable communications are identified as foundational enablers of these sustainability objectives. In addition, high data rates are recognized as essential in specific scenarios, particularly within consumer subnetworks, to support advanced multimedia applications and real-time interactivity. Related deliverable: D2.1, “Initial definition of scenarios, use cases and service requirements for in-X subnetworks” , D2.2, “Refined definition of scenarios, use cases and service requirements for inX subnetworks”. 2.1 Architectural enablers In-X subnetworks, along with their integration in the 6G network of networks, represent a major architectural innovation in the wireless communication landscape. In-X subnetworks are specialized networks designed to provide localized, high-performance 16 5. RADIO RESOURCE MANAGEMENT Subnetworks can spontaneously become very dense, such as those installed in robots, production modules, and vehicles. Interference is therefore a major challenge to be dealing with in order to support the demanding requirements of dense subnetworks. Significant progress has been achieved in advancing reliability, scalability, low latency, spectral efficiency, and resilience to interference through a unified suite of centralized, distributed, and goal-oriented radio resource management (RRM) solutions developed within the 6G-SHINE project. Figure 6 presents a pictorial depiction of the RRM context for dense in-X subnetworks. Centralized solutions rely on the presence of RRM capabilities at the 6G parent network. Subnetworks report periodically their experienced channel conditions to the parent network, that can perform decisions on the radio resources they should use locally. Approaches based on deep neural networks (DNN) allow to reduce the computational burden with respect to iterative heuristics in the execution phase, provided that the DNN has been previously trained in the environment of interest. To mitigate the effects of outdated channel state information (CSI) due to subnetwork mobility, and combining spatio-temporal attention-based LSTM prediction with a robust deep neural network (DNN)-based resource allocation scheme. Under realistic CSI delays (e.g., 4-samples lag), this method outperformed state-of-the-art approaches, delivering 53% higher minimum spectral efficiency than systems without prediction and 94% higher than those using standard LSTM predictors. These results were validated in extremely dense Figure 6. Radio resource management in dense in-X subnetworks. 17 scenarios—up to 25,000 subnetworks per km²—supporting the project's goal of scaling 10× beyond current 5G ultra-dense deployments. A complementary distributed RRM solution tailored to subnetworks that are not in the coverage area of a parent network, based on Graph Neural Networks (GNNs), enables autonomous power control in environments where centralized coordination is not feasible. This approach improved spectral efficiency by ~7% in uniform settings and up to 13.16% in heterogeneous channel conditions over equal power allocation baselines, while leveraging practical over-the-air message exchanges compatible with existing 3GPP systems. For mission-critical applications in industrial automation, a goal-oriented RRM strategy was introduced, co-optimizing network quality and application-level metrics such as robot mission completion time. Using a Proximal Policy Optimization (PPO) reinforcement learning algorithm for mobility control, the system increased the likelihood of meeting stringent URLLC targets—achieving 20% higher probability of sustaining the same BLER under a 0.5 ms latency constraint compared to baseline approaches. In light of the rising relevance of unlicensed and hybrid spectrum use, 6G-SHINE developed techniques to enable dense subnetwork operation in such environments. Semi-static channel access methods were shown to support up to 10× more XR sessions than conventional dynamic schemes. When combined with licensed-assisted operation, an additional 67% increase in supported subnetworks was achieved. Further capacity gains—up to 40%—were realized by mitigating in-band emissions through improvements in device front-end design and spectrum coordination strategies. To address the pervasive issue of external interference, a dual-pronged strategy was employed: robust resource allocation and resilient receiver design. A Gradient Descentbased Resource Allocation (GDRA) algorithm limited spectral efficiency degradation to just 9.7%, outperforming existing benchmarks that saw losses exceeding 13%. On the receiver side, low-complexity likelihood ratio approximation techniques were developed to ensure reliable decoding even under impulsive noise and jamming, with dynamic adaptation to real-time channel conditions and minimal processing burden. Together, these innovations demonstrate the feasibility of deploying dense, autonomous, and interference-resilient subnetworks across diverse domains—including industrial, consumer, and vehicular settings. The RRM solutions developed not only exceed current performance benchmarks but also provide a robust foundation for enabling dynamic and scalable 6G network architectures. Related deliverables: D4.1 “Preliminary results on the management of radio resources in subnetworks in the presence of legitimate and malicious interferers”, D4.3 “Final results on the management of radio resources in subnetworks in the presence of legitimate and malicious interferers”. 18 6. MANAGEMENT OF TRAFFIC, SPECTRUM AND COMPUTATIONAL RESOURCES Recent developments have addressed the management of traffic, computation, and spectrum resources both among subnetworks within the same domain and between subnetworks and the broader 6G infrastructure. These advancements have built upon earlier architectural frameworks to define a comprehensive operational model for subnetworks, prioritizing user privacy, autonomy, and reduced dependency on the central network. A new category of user equipment has been introduced, with tailored enhancements in configuration, security, and data multiplexing. A decentralized authentication framework supports direct device-to-device interactions, minimizing reliance on core network functions. Subnetwork formation, registration, and mobility management have been further refined through distributed control and user-plane functions, with added support for local IP routing. To optimize control plane operations, especially for tasks like location updates, offloading strategies were proposed. Subnetworks were also enabled to coordinate with neighboring subnetworks, supporting predictive mobility and coordinated measurement functions, ultimately improving individual user performance. The challenges of handling multi-modal data—especially relevant in consumer applications—were also addressed. Two methods were developed to align data packets across interrelated flows, either by enriching packet headers for device-side synchronization or by introducing a management-side controller that handles scheduling and alignment. These methods enhance quality of service even for locally-contained traffic. Improvements in uplink scheduling were also proposed, enabling more efficient resource use by the parent network. In computation offloading, the framework now supports both local and decentralized approaches across subnetworks. Methods for selecting compute nodes and managing the association between operational and computational entities have been formalized. To support this shift, the quality of service framework was expanded to incorporate computing-specific requirements and constraints, with new parameters and procedures introduced for supporting computation-aware service levels. In vehicular applications—where topologies are relatively static, but reliability requirements are extremely stringent—novel deterministic task offloading and resource allocation schemes were proposed. These prioritize task deadlines rather than mere latency minimization, allowing for better workload distribution across IoT-edge-cloud architectures. This improves scalability and performance in safety-critical systems. Further enhancements for in-vehicle networks include deterministic task scheduling schemes tailored for zonal electronic/electric architectures, enabling reliable service levels even under increasing computational demands. For immersive XR applications, a computeand network-aware traffic steering framework was introduced. It dynamically selects optimal service instances and manages mobility-aware service anchoring and migration, ensuring seamless execution in heterogeneous and resource-constrained environments, thereby enhancing both user experience and quality of service. 19 Finally, research on dynamic spectrum sharing compared regulatory policies across major global regions, evaluated existing sharing mechanisms, and highlighted emerging trends. A flexible spectrum access protocol was developed to allow dynamic allocation of licensed resources to subnetworks based on real-time traffic demands, using the concept of dynamic resource pools to enhance spectral efficiency and responsiveness. Related deliverables: D4.2 “Preliminary results on the management of traffic, computational and spectrum resources among subnetworks in the same entity, and between subnetworks and 6G network”, D4.4 “ Final results on the management of traffic, computational and spectrum resources among subnetworks in the same entity, and between subnetworks and 6G network”. 7. DEVELOPED PROOF-OF-CONCEPTS A selected set of technologies developed within the project were demonstrated through proof-of-concepts (PoCs) in laboratory facilities. Since 6G-SHINE is a low technology readiness level (TRL) project, it was out of our scope to develop a fully integrated prototype of an in-X subnetwork. Instead, the focus was on showcasing specific technology components, selected on the basis of their promised performance and feasibility of implementation within the project time. The demonstrated PoCs are the following: 1. Low-latency channel emulator – Validates short-range channel modeling for vehicular, industrial, and consumer applications, significantly improving the latency performance of the PROPSIM channel emulator in capturing short-range channel effects. 2. Latency-aware MAC access – Demonstrates improved coexistence in shared bands, providing deterministic medium access for latency-sensitive applications, with little to no penalization of best-effort traffic. 3. Jamming-resilient PHY – Shows how an enhanced scheduler and decoder can maintain robust performance under both wideband and narrowband jamming conditions. 4. Intra-subnetwork macro-diversity – Uses network coding and cooperative mechanisms to improve reliability and latency in industrial subnetworks. 5. Centralized RRM – Introduces dynamic resource allocation mechanisms for interference mitigation in vehicular and industrial environments. 6. Centralized/distributed interference management – Demonstrates AI-driven, decentralized techniques to optimize transmission power and channel usage in subnetworks operating over overlapping resources. 7. Jamming detection via anomaly detection – Presents anomaly detection solutions capable of recognizing malicious jammers and thereby triggering appropriate mitigation techniques in subnetworks. These PoCs confirmed that the technologies developed in the project can deliver performance gains comparable to those observed in theoretical and simulation studies, 20 while accounting for the constraints of the laboratory environment and the underlying hardware and software. Related deliverables: D5.2 “First implementation and evaluation of PoCs”, D5.3 “Second implementation and evaluation of PoCs, final results”. 8. OUTLOOK The 6G-SHINE project has significantly advanced the vision of in-X subnetworks, shortrange low-power cells at the very edge of the 6G “network of networks.” Scientifically, the project has contributed new models, architectures, and methods across the physical, MAC, and resource management layers. These results are the building blocks for making in-X subnetworks (and their integration with a 6G parent network) a solid reality in the figure wireless landscape, and form a foundation for future research on resilient, efficient, and intelligent short-range wireless systems. From an industrial and economic perspective, 6G-SHINE has already left a tangible mark on the standardization landscape. The project introduced the subnetwork concept in 3GPP, contributed to ETSI ISAC, and delivered inputs to IEEE and ETSI RIS and THz groups. In particular, our proposed in-vehicle subnetwork use case has been already captured on 3GPP TR 22.870 “Study on 6G Use Cases and Service Requirements”, section 11.10. This early engagement ensures that the technologies and use cases explored in 6G-SHINE can influence global standards at the very moment when 6G discussions are consolidating. The consortium also generated substantial intellectual property and initiated collaborations that strengthen Europe’s industrial competitiveness in wireless technologies. Societally, the project promotes a sustainable and human-centric approach to 6G, with localized processing for privacy and safety, reduced cabling for greener deployments, and support for new applications in education, industry, and mobility. Beyond its technical outcomes, 6G-SHINE has also contributed to the broader SNS and 6G-IA vision activities, ensuring that the project’s innovations are visible and aligned with Europe’s strategic ambitions. In summary, 6G-SHINE has positioned in-X subnetworks as a credible building block of 6G, bridging low-TRL research with standardization and paving the way for their industrial and societal uptake.