Programa del congreso

This paper presents a comparative analysis and propagation modeling of path loss (PL) in terahertz (THz) channels for indoor scenarios at 300 GHz. Experimental measurements in a conference room are used to evaluate several literature-based PL models by comparing their root-mean-square error performance. A custom-developed three-dimensional ray-launching simulation tool is introduced and validated against measurement data, yielding a mean absolute error of 0.77 dB. Furthermore, the study examines the impact of receiver height on PL characteristics by applying the Close-In PL model across multiple vertical levels of the conference room, revealing significant variations in the PL exponent and standard deviation. These findings underscore the need for height-specific channel models to support the design of robust THz communication systems for next-generation networks.

Intelligent systems increasingly leverage contextual data, such as air quality or user location, to provide enriched, context-aware services. The widespread use of IoT devices, along with advancements in computational capabilities and deep learning, is driving the development of more complex systems supporting diverse applications. However, these context-aware environments present challenges related to interoperability, scalability, and device/platform heterogeneity. Addressing them requires a structured approach, defining a set of software entities running across different contextual units and information systems that function as data collectors and aggregators. These units may employ from simple decision rules to state-of-the-art machine learning techniques to acquire knowledge and can potentially operate as distributed intelligent agents. A critical aspect of such systems is secure communication between these software entities: robust authentication, integrity, and confidentiality mechanisms are essential to maintaining system trust. This article explores the challenges of secure communications in complex, agent-based, context-aware scenarios. Moreover, it also reviews technologies to ensure confidentiality, integrity, and authentication.

The integration of cognitive systems and ambient intelligence has led to significant advancements in user-adaptive environments, particularly within Ambient Assisted Living applications. Audio-based interactions provide an intuitive and efficient means of communication, enabling a broad range of cognitive and assistive services. However, these advancements introduce substantial security and privacy concerns, particularly when processing and storing the associated data. This article presents a preliminary approach to audio-based authentication, utilizing machine learning techniques to enhance security in cognitive environments by ensuring accurate identification of individuals based on voice fingerprinting. Moreover, it outlines a data protection strategy and discusses privacy-preserving mechanisms to ensure compliance with ethical and regulatory standards.

Este trabajo presenta una solución innovadora basada en la computación Edge y la inteligencia artificial para detectar en tiempo real conductas peligrosas en entornos urbanos multimodales. Las imágenes son procesadas localmente en dispositivos embebidos desplegados en la infraestructura urbana, evitando la transmisión de imágenes sensibles y garantizando plenamente la privacidad de peatones y vehículos según las normativas europeas y españolas (RGPD y LOPDGDD). Mediante modelos avanzados de aprendizaje profundo y técnicas robustas de seguimiento multiobjeto, el sistema genera analíticas anonimizadas y alertas de seguridad proactivas. La solución propuesta contribuye significativamente a mejorar la seguridad vial, la privacidad ciudadana y la gestión inteligente del tráfico dentro del paradigma de ciudades cognitivas.

This paper presents an approach for the periodic collection of environmental data using an autonomous heterogeneous fleet of Automated Guided Vehicles (AGVs). The system integrates SLAM-based navigation, a decentralized architecture leveraging ROS 2, and secure communication protocols to ensure resilience in cognitive environments. A deployment in a real-world university setting is discussed, highlighting system performance and validation results. The proposed solution enhances se-

cure data acquisition while maintaining adaptability to dynamic environments. Additionally, its modular and scalable design makes it particularly well-suited for cognitive environments, such as smart campuses or industrial testbeds, where autonomous perception, decision-making, and interoperability are critical.

This paper focuses on the importance of radio planning and assessment of industrial environments for the future Cognitive Industry and Industry 5.0. The current and most commonly used wireless communication technologies, such as those based on the standard IEEE 802.11, Bluetooth, ZigBee, UWB, LoRaWAN, SigFox or NB-IoT already pose significant challenges regarding coexistence, coverage/capacity and noise, and interference immunity in industrial environments. The implantation of new rising technologies (e.g., Mioty, WiFi HaLow), the deployment of 5G networks in such environments (for the use of complete 5G services or RedCap devices) and the new frequency bands coming with 6G networks will create much more complex industrial environments in terms of radio planning tasks. Considering that, this work presents a deterministic simulation tool for volumetric RF analysis of industrial environments, showing the relevance of such accurate tools for the development of the future cognitive industry based on reliable and resilient wireless communication systems.

Wearable devices have gained significant popularity recently due to their ability to collect vast amounts of physiological data non-intrusively. However, these devices' resource-constrained nature often prioritises their operations' efficiency while security considerations are overlooked. This paper identifies the security vulnerabilities of EmotiBit bracelets, an open-source wearable device commonly used in healthcare and education projects. To address concerns related to data confidentiality and device authentication within this ecosystem, we propose two security schemes based on pre-shared keys and public-key cryptography. Additionally, the paper discusses the details of the implementation of these schemes and provides an evaluation of their strengths and limitations.