i2CAT

This keynote presents a bold, transformative vision for the future of mobile connected systems based on Agentic AI. We will delve into our research on leveraging Small and Large Language Models (SLMs and LLMs) to create a new generation of intelligent, autonomous mobile networks. As a foundational proof-of-concept, we will introduce "TelecomRAG," a novel framework using Retrieval-Augmented Generation that enables LLMs to understand and reason over complex telecommunication standards with high precision and verifiability. Building on this, the talk will explore how to unlock the disruptive potential of Agentic AI to fundamentally reshape the design, deployment, and evolution of future connected systems.

Development of maneuver cooperation strategies is essential for enhancing traffic safety and efficiency. In a Vehicle-To-Everything (V2X) ecosystem, Maneuver Coordination Messages (MCM) allow road users to coordinate trajectory intentions based on decision-making mechanisms so they can effectively solve complex interactions such as unsignalized intersection crossings or lane changes. As these mechanisms are still in the early stages of development, research continues exploring their effectiveness and feasibility. In this paper, a Decentralized Maneuver Coordination Service (D-MCS) implementation is evaluated in real-world conditions where Vulnerable Road Users (VRUs) and emergency vehicles are taken into account, assessing the efficiency of the cooperative algorithm by demonstrating a reduction in execution time to complete the maneuvers. The results of the experiments also demonstrate that the proposed DMCS achieves a significant reduction in maneuver coordination time compared to a centralized MCS and improves traffic flow and VRU safety by allowing vehicles to proactively adjust their maneuvers.

Digital twins are living digital surrogates of objects and processes from the world around us and serve many purposes in situations where a physical model cannot be used. Digital twins can be considered to still be in their nascent phase, be it in the variety of application domains or in the science of creating and manipulating them. The DIDYMOS-XR project uses digital twins to create large-scale (city-wide) and long-term eXtended Reality (XR) applications. In this paper, we propose three use cases—city planning, city maintenance, and city tourism—along with an industrial use case for manufacturing environments, demonstrated through evaluation sites in Germany and Spain. The paper introduces cutting-edge solutions in the creation, simulation, and manipulation of DTs, including advancements in 3D scene reconstruction, data enhancement, data compression, sensor fusion, localization, rendering, and scene understanding. Additionally, we address the ethical and privacy challenges of digital twin systems and propose strategies for mitigating these issues. This work offers a comprehensive framework for creating semantically rich, scalable, and interactive digital twins, providing valuable insights for diverse applications in urban planning, maintenance, tourism, and industrial optimization.

The Service-Based Architecture (SBA) has been successfully employed in the 5G Core since Release 15. Its modular and loosely coupled approach to network function interaction, based on the producer/consumer paradigm, perfectly pairs with the softwarization trend mobile networks have undergone over the last decade. Still, the Radio Access Network (RAN) did not leverage this approach, retaining the traditional Protocol-Based Architecture (PBA). In this paper, we make the case for adopting the SBA in the RAN, a transition we argue is both advantageous and feasible. We first present a comprehensive analysis of the benefits, such as enhanced flexibility and innovation, while also considering potential implementation challenges. Then, we detail how the SBA can be used to provide two exemplary procedures currently performed using the PBA, showing the feasibility of a fully service-based 5G network architecture.

Social Virtual Reality (VR) enables shared media experiences between remote people inside immersive and realistic 3D spaces, providing richer and more natural interactions than in classical 2D social conferencing tools. Likewise, the benefits and engagement can even be magnified by integrating realistic and volumetric user representations (i.e., 3D holograms) in these virtual environments rather than synthetic avatars. This paper presents the design and evaluation of an interactive Social VR scenario for a joint and collaborative exploration of a catalogue of professional video clips by a broadcaster. On the one hand, the scenario includes a control panel to select the desired year and clip. After the year selection, a time travel through a lift effect is enforced to teleport users through a multi-level semi-open building in which each level / floor represents one year, and its look-and-feel is customized to resemble that year. On the other hand, the scenario allows the integration of up to four users represented as 3D holograms (full-body and full volume Point Clouds), each one with his/her own screen for video consumption, and arranged in a cross 360<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∘</sup> shape to allow for a natural visual interaction among themselves. The evaluation results with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N=48</i> professionals of the broadcast sector not only provide relevant insights about the technical requirements and obtained performance, but confirm the satisfactory user experience (in terms of presence, togetherness, quality of interaction) provided by the presented technology and VR scenario and, most importantly, reveal and contribute to identifying the potential and opportunities of Social VR in the broadcast / video consumption landscape.

The increasing complexity of Extended Reality (XR) applications demands substantial processing power and high bandwidth communications, often unavailable on lightweight devices. Remote rendering consists of offloading processing tasks to a remote node with a powerful GPU, delivering the rendered content to the end device. The delivery is usually performed through popular streaming protocols such as Web Real-Time Communications (WebRTC), offering a data channel for interactions, or Dynamic Adaptive Streaming over HTTP (DASH), better suitable for scalability. Moreover, new streaming protocols based on QUIC are emerging as potential replacements for WebRTC and DASH and offer benefits like connection migration, stream multiplexing and multipath delivery. This work describes the integration of the two most popular multimedia frameworks, GStreamer and FFmpeg, with a rendering engine acting as a Remote Renderer, and analyzes their performance when offering different protocols for delivering the rendered content to the end device over WIFI or 5G. This solution constitutes a beyond state-of-the-art testbed to conduct cutting-edge research in the XR field.

In the rapidly evolving landscape of the Internet of Medical Things (IoMT), ensuring the security and reliability of interconnected medical devices is paramount. Traditional cybersecurity methods, while extensively discussed, often fall short in detecting and responding to evolving threats. Anomaly-based Intrusion Detection Systems (IDS) leveraging Artificial Intelligence (AI) techniques are increasingly favored for their ability to detect zero-day attacks and adapt to new threat landscapes. Accordingly, this doctoral thesis aims to develop and validate AI-based IDS for identifying and mitigating cybersecurity threats within IoMT environments. By utilizing publicly available datasets and generating a novel dataset from real-world IoMT devices and simulated cyberattacks, this study enhances detection system performance and investigates the generalizability of the proposed IDS across diverse environments using innovative AI techniques. Additionally, it explores and designs effective automated prevention techniques to enhance patient-centric cybersecurity against potential threats. The findings of this research are expected to significantly contribute to IoMT security, offering practical solutions for protecting critical healthcare infrastructure and ensuring patient safety.