Scientific publications
See the scientific impact of our work. Discover the scientific papers supported by COP-PILOT
Abstract
6G network complexity necessitates high levels of autonomy, yet current intent-based systems struggle with ambiguous or incomplete human requests. This paper introduces an agent-based, intent-driven end-to-end (E2E) orchestration framework designed for Network-as-a-Service (NaaS) delivery through collaborative intent co-creation. The proposed system leverages a pool of Domain Expert Agents and a TM Forum-aligned Body-of-Knowledge (BoK) to iteratively refine user requests into deterministic, machine-readable actions. A fundamental design principle is the decoupling of cognition and actuation, where AI-driven reasoning is isolated from standardized execution controllers to ensure safety and operational trust. The framework includes a dual-layer memory system to maintain coherence during multi-step collaborations. The presented prototype, built on ETSI OpenSlice and the Model Context Protocol (MCP), evaluates across several open-source Large Language Models (LLMs). While these models demonstrate high instruction compliance, results reveal a significant gap in translating high-resolution intents into valid, catalog-backed orders without hallucinations.
Relevance of the Paper to the COP-PILOT Project:
The concept of intent refinement can be effectively applied to the COP-PILOT architectural components. In particular, the COP-PILOT Business Portal is designed to capture stakeholders’ intent and translate it into appropriate specifications for the End-to-End Service Orchestrator. This approach is especially valuable in complex scenarios where a single intent must be mapped to multiple specifications.
Authors
Kostis Trantzas, Besiana Agko, Christos Tranoris, Irene Denazi
Abstract
This work proposes an agentic, intent-driven end-to-end (E2E) orchestration framework that integrates intent co-creation with a Test-Driven Quality Assurance paradigm. In this framework, autonomous agents iteratively refine a user’s initial intent into a confirmed, auditable specification. Furthermore, the system automatically derives validation tests from these intents before provisioning, directly mirroring the Test-Driven Development workflow in software engineering to ensure proactive Service Level Agreement (SLA) compliance. The architecture is grounded in a standards-aligned knowledge representation using TM Forum (TMF) information models and catalogs. This enables deterministic graph traversal from high-level Product Offerings down to granular Service/Resource and Test specifications. We prototyped this architecture by extending OpenSlice with a message-driven, multi-agent pattern and integrating MCP-enabled (Model Context Protocol) tool access for real-time knowledge retrieval. Currently, our evaluation of the agents targets the intent co-creation phase as a baseline toward full-scale orchestration. Building on experiments with multiple open-source Large Language Model (LLM) backends integrated with the TMF-based knowledge base, we observe substantial variability in tool-use reliability and hallucination patterns, underscoring the critical importance of robust knowledge integration in agentic 6G systems.
Relevance of the Paper to the COP-PILOT Project:
The concepts of intent refinement and test-driven quality assurance presented in this paper can be effectively applied to the COP-PILOT architectural components. In particular, the COP-PILOT Business Portal aims to capture stakeholders’ intent and translate it into appropriate specifications for the End-to-End Service Orchestrator. These approaches are especially valuable in complex scenarios where a single intent must be mapped to multiple specifications. Additionally, when supported, test-driven quality assurance can enhance the Domain Orchestration layer by enabling systematic validation of deployed platform services.
Authors
Christos Tranoris, Besiana Agko, Kostis Trantzas, Irene Denazi
Abstract
Relevance of the Paper to the COP-PILOT Project:
Authors
Maria Crespo-Aguado, Lucía Martínez-Palomo, Nuria Molner, Arturo-José Torrealba-Ferrer, Jose-Miguel Higón-Sorribes, Carlos Blasco, Carlos Ravelo and David Gomez-Barquero
Abstract
The monitoring of oxygen in food packaging during storage and transportation is crucial in food quality surveillance, warning users regarding food spoiling, happening through compound oxidation and aerobic microorganism proliferation. In this overall scenario, we report the development of a
flexible, cost-effective, and Bluetooth-assisted electrochemical sensor for oxygen detection in food packaging. The device encompasses three layers, namely a zinc sheet as an anode, a conductive silver
ink printed on an oriented polypropylene sheet serving as a cathode, and a deep eutectic solvent deposited on a paper-based substrate sandwiched between both electrodes. The sensing tool provided a wide linear range for oxygen detection up to 20.9 O 2 % v/v with good intra-electrode
repeatability (RSD % = 0.02 %). Finally, the developed device was integrated with a 3D printed holder and tested for oxygen detection in packages containing mushrooms, tomatoes, and broccoli samples,
obtaining a good correlation with the reference method. This study opens noticeable possibilities for employing paper-based metal-air batteries in the detection of specific target analytes, by integrating
paper substrate and metal-based batteries delivering smart and self-powered instruments as reliable and accurate analytical tools.
Relevance of the Paper to the COP-PILOT Project:
The study carried out in this paper provides valuable insight into the use of Bluetooth-based and wireless sensing platforms. Specifically, this technology will be utilized in the framework of the COP-PILOT project to develop a Bluetooth-assisted electrochemical sensor for plant health monitoring.
Furthermore, with the investigation performed during this work, we dealt with the agrifood field, and specifically with the monitoring of parameters related to vegetables, laying the groundwork for the development of a wearable electrochemical sensor for monitoring (anti)nutrients in plants. This is strictly related to the COP-PILOT project, following the Use Cases, and to fulfill the Expected Outcomes related to Cluster 3A.
This is an open access article under the CC BY license http://creativecommons.org/
Authors
Achref Chebil, Vincenzo Mazzaracchio, Leonardo Duranti, Ludovica Gullo, Fabiana Arduini
Model Optimization pipeline for hardware-aware trustworthy edge- residing AI: a grid reliability case study
Abstract
This paper presents a hardware-aware optimization pipeline for deploying trustworthy edge AI on resource-constrained devices, demonstrated through an automated fault inspection system for electrical insulators in grid infrastructure. The approach combines flat minima training methods with hardware-specific quantization and a hierarchical inference paradigm, where a calibrated far-edge model handles high-confidence classifications on-device while offloading uncertain samples to a more powerful edge server. Experimental results across multiple embedded platforms demonstrate 98% cooperative accuracy with power consumption as low as 11.5 mJ per inference, supporting reliable real-time operation in safety-critical energy environments.
Relevance of the Paper to the COP-PILOT Project:
The paper addresses automated fault inspection of electrical insulators running on resource-constrained far-edge devices — a challenge that maps directly onto Cluster 3E’s core mission of enhancing energy system resilience and efficiency through edge intelligence deployed across distributed grid infrastructure in Western Greece.
Cluster 3E explicitly targets the transition from reactive to predictive maintenance paradigms across all three use cases (UC#3E.1–3E.3). The paper’s insulator fault detection system embodies precisely this transition: rather than waiting for infrastructure failures, it enables real-time, on-device classification of faults before they escalate.
Authors
Alexandros Machairas, Nikolaos Tzanis, Athanasios Bachoumis
Abstract
In 2023, the World Economic Forum selected wearable plant sensors as one of the Top 10 Emerging Technologies, demonstrating that these smart analytical tools will be relevant in the next generation of agrifood practices. Considering the robustness, accuracy, and miniaturisation of electrochemical (bio)sensing tools, electrochemical-based plant sensors could be suitable devices to address the requirements for their advanced applications in the agrifood sector. This review deals with electrochemical (bio)sensors for monitoring agrochemicals , phytohormones, growth precursors, and stress biomarkers, using wearable and implantable configurations. The design and type of biocomponent and/or nanomaterial(s) used are reported, highlighting the analytical performances obtained on plants. The ongoing application of these analytical tools is discussed, and the future applications combined with Internet of Thing and Artificial Intelligence are envisioned, with the overriding aim to give an overall scenario related to plant electrochemical (bio)sensors for the next technologies in the agrifood sector.
Relevance of the Paper to the COP-PILOT Project:
This review paper focuses on the state-of-the-art in wearable and implantable electrochemical biosensors for monitoring plant health, including agrochemicals, phytohormones, and stress biomarkers. The insights from this research are crucial for one of COP-PILOT’s key objectives in Cluster 3A, which is to develop a wearable electrochemical sensor for monitoring antinutrients in leaves. The study provides essential knowledge for the development of our sensing device, including its implementation and integration with our cluster partners, enabling us to create a reliable and innovative analytical tool for plant health monitoring.
This is an open access article under the CC BY license http://creativecommons.org/
Authors
Narjiss Seddaoui, Fabiana Arduini
Abstract
Relevance of the Paper to the COP-PILOT Project:
–
Authors
Johan Kristiansson; Ulf Bodin; Carl Borngrund; Jerker Delsing; Jesper Martinsson
Relevance of the Paper to the COP-PILOT Project:
The paper focuses on developing an analytical platform that combines sample treatment with an electrochemical biosensor to measure phytic acid in spinach leaves. It strictly aligns with the COP-PILOT project (Cluster 3A), providing insights into creating a sensor to quantify a nutritional marker for assessing crop health and quality (Use Cases #3A). Additionally, the use of a wireless portable system enables integration with IoT technology for real-time environmental and crop health monitoring.
Authors
Ludovica Gullo, Igor Gabriel Silva Oliveira, Achref Chebil, Luca Fiore, Silvia Maria Martelli, Willyam Róger Padilha Barros, Fabiana Arduini
Abstract
Abstract—Cyber-Physical Systems (CPS) powered by Artificial Intelligence (AI) have the potential to revolutionize industries by enabling advanced analytics and autonomous decision-making. To support resource-intensive applications, there is often a need to dynamically allocate additional compute resources. The Edge-Cloud Continuum enables allocation and deployment of workloads across platforms, including IoT devices, edge clusters, and cloud environments. However, the growing computational demands of these systems can unfortunately result in increased energy consumption and higher carbon emissions.
This paper investigates the development of a carbon-aware scheduler for the Edge-Cloud Continuum, designed to optimize workload placement by balancing energy consumption, temporal variations in carbon intensity, and resource availability. Key contributions of the paper include a spatiotemporal scheduling algorithm, a discrete-event simulator capable of replaying realistic workloads from the MIT SuperCloud dataset, and a comprehensive empirical evaluation.
Findings from the paper show substantial reductions in carbon emissions by prioritizing renewable energy sources and time shifting workloads to periods of lower carbon intensity. However, when clusters operate under high utilization, time-shifting can inadvertently result in significantly higher emissions. In such scenarios, simpler greedy algorithms can be more effective.
Relevance of the Paper to the COP-PILOT Project:
–
Publication
- 30 July 2025
Authors
Johan Kristiansson, Jerker Delsing, Thomas Ohlson Timoudas
Abstract
Cloud-native computing has transformed modern application development, deployment, and management by enabling scalability and flexibility. However, the increasing complexity of workloads and the dynamic resource demands challenge traditional scheduling and resource provisioning techniques, often leading to inefficiencies. This paper explores AI-driven approaches to optimizing cloud-native scheduling and resource provisioning. By leveraging machine learning, deep reinforcement learning, and predictive analytics, AI enhances decision-making, automates scaling, and improves workload distribution. We present a comprehensive review of recent AI techniques applied to container orchestration and Kubernetes-based scheduling, analysing their impact on cost
reduction, performance optimization, and resource efficiency. Additionally, we discuss key challenges
such as model interpretability, real-time adaptability, and integration with existing cloud and edge infrastructures. Ultimately, this paper provides insights into the future of intelligent cloud and edge resource management, emphasizing the necessity of AI-augmented strategies to meet the growing demands of next-generation applications.
Relevance of the Paper to the COP-PILOT Project:
The paper provides a comprehensive survey of how artificial intelligence (AI) is being integrated into cloud-native systems, highlighting its role in enhancing automation, scalability, and decision-making in complex distributed environments. This paper helps COP-PILOT gain insight into current trends, key challenges, and emerging research directions for applying AI to optimise cloud-native, multi-domain architectures.
Publication
- 23 June 2025
Authors
Tomás Dias, Luís Ferreira, Diogo Fevereiro, Luis Rosa
