Publications

This paper examines the COMMECT project's initiatives to overcome the digital divide in rural areas through the deployment of integrated terrestrial (cellular XG, public and private 5G networks, Internet of Things, NB-IoT, LoRa, etc.) and Non-Terrestrial (satellites, drones) Networks. The project has established five user-centric Living Labs across Europe, focused on viticulture (Luxembourg), forestry (Norway), livestock transport (Denmark), olive farming (Türkiye), and sustainable agriculture (Serbia). To facilitate connectivity decision-making, the project is developing a Decision-Making Support Tool (DST), which leverages Large Language Models (LLMs) to provide tailored connectivity recommendations. This paper outlines the project's approach, the integrated networks deployed in the Living Labs, and the DST’s functionalities and future developments.

A lot of attention has traditionally been given to digitalization of smart cities and urban areas. However, bridging the digital dive for the rural areas has become necessary to exploit the full economic growth from new technologies since many industries also belong to the countryside. This paper introduces novel insights into use cases from the forest industry and the perceived impact from innovations activated from technologies such as the fifth-generation mobile technology (5G) and Internet of Things (IoT). An agile methodology is followed developing the forest use cases to handle the risk and uncertainty within this domain. After interviews with stakeholders from the forest value chain we found that digitalization of the forest value chain is necessary. We also found three distinct use cases with different scenarios where premium cellular network (5G) connection can make an impact. The potential impact dimensions were identified and operationalized in three groups: business, user/society, and environment. Our findings will set the scene for the forthcoming specification of the use case requirements and the validation of Proof of Concept’s (PoC’s) during the next step in our agile methodology. Our work is funded through Horizon Europe program COMMECT and PoC’s trial results from the Norwegian forest living lab will be published at a later stage.

The next generation network technology 5G introduces new business opportunities for telecom operators that can help them mitigate their declining revenues. 5G reduce the network latency and enables a massive number of Internet of Things (IoT) devices to be connected to its network. This paper proposes a co-creative approach for telecom operators when they exploit 5G/IoT technologies in industry use cases in rural communities and builds upon the concept of collaborative business modelling. Partnering with providers of digital niche services and in-depth understanding of users’ pains and challenges is key when establishing “beyond connectivity” solutions. Moreover, the collaborative model secure long-term survivability in which all ecosystem stakeholders benefit through participation, hence moving away from individual profit maximization. The initial design of the collaborative business model occurs as a part of an agile development process, in which stakeholder needs, and requirement are uncovered together with technical trials and commercial validation of suggested solutions (MVP's/protocepts). Our findings provide novel insights towards the understanding of how collaborative business models support the deployment of enhanced connectivity enabled solutions (remote operation support and fire situation awareness service) in the forestry industry in rural areas of Norway, as well as the role of the telecom operator in orchestrating the ecosystem. Updated business models based on forthcoming development and trials of these two solutions will be presented in later publications.

The use of 5G private networks is gaining popularity in the industry due to their capability to address the requirements of a large number of vertical markets including own administration and communication privacy and security. These locally administrated networks are similar to micro-operators, providing low delay, high capacity, and reliable communication to the local devices. However, these networks will need to interact with each other in order to extend coverage and provide connectivity to visiting devices. In this article, a roaming framework addressing geographically distributed private networks is proposed, by adapting the 5G roaming system of the large-scale operators to a federated authentication and authorization framework. To realize the proposed solution, the Fraunhofer FOKUS Open5GCore toolkit is extended with the proposed architectural extension and it is then evaluated on the basis of flexibility and privacy of deployment, backhaul usage, and reduced network administration underlining its feasibility for further development.

Validating the impact from enhanced connectivity solutions has frequently been related to technological dimensions such as speed, capacity, and coverage. Moreover, research efforts on adoption of digital technologies have traditionally been dedicated to communities in cities and urban areas. Lately, more attention has been paid to socio-economic and environmental dimensions to secure that new digital innovations introduced for the market are business and community friendly as well as environmentally sustainable. Our hypothesis is that connectivity solutions enhanced by 5G and IoT will have large impact on rural communities and industries prevalent in rural areas. The backdrop is that agricultural and forest-based businesses are typically found in rural areas and have just started their digitalization journeys, meaning many benefits are still left uncaptured. This paper suggests a methodology for validating societal and economic impact of new connectivity solutions in rural areas, building upon a macro and micro level analysis. The former assesses effects from increased broadband coverage on demographical outcomes on a regional/municipality level, whereas the latter assesses the effects on individual stakeholders within respective rural areas, focusing on concrete deployment of use cases. The solutions are designed based on elicited user needs and requirements, and our agile approach helps us transferring these insights into the design of the solutions that later will be tested in small- and large-scale trials. Preliminary findings indicate a significant relationship between increased broadband connectivity and social outcomes on a regional level and display an array of expected impact indicators to be captured by stakeholders on a use case level. The proposed methodology offers novel insights into the steps of validating the impact of next generation digital technologies in rural areas. It complements existing work on the validation of smart connectivity solutions and that traditionally has been given to cities and urban areas.

Validating the impact from enhanced connectivity solutions has frequently been related to technological dimensions such as speed, capacity, and coverage. Moreover, research efforts on adoption of digital technologies have traditionally been dedicated to communities in cities and urban areas. Lately, more attention has been paid to socio-economic and environmental dimensions to secure that new digital innovations introduced for the market are business and community friendly as well as environmentally sustainable. Our hypothesis is that connectivity solutions enhanced by 5G and IoT will have large impact on rural communities and industries prevalent in rural areas. The backdrop is that agricultural and forest-based businesses are typically found in rural areas and have just started their digitalization journeys, meaning many benefits are still left uncaptured.

We present a method for analyzing character errors for use with character-based, end-to-end ASR systems, as used hereinfor investigating dialectal speech. As end-to-end systems are able to produce novelspellings, there exists a possibility that the spelling variants produced by these systems can capture phonological information beyond the intended target word. We therefore first introduce a way of guaranteeing that similar words and characters are paired during alignment, thus ensuring that any resulting analysis of character errors is founded on sound substitutions.Then, from such a careful character alignment, we find trends in system-generated spellings that align with known phonological features of Norwegian dialects, in particular, “r” and “l” confusability and voiceless stop lenition. Through this analysis, we demonstrate that cues from acoustic dialectal features can influence the output of an end-to-end ASR systems.

Accurately forecasting traffic in telecommunication networks is essential for operators to efficiently allocate resources, provide better services, and save energy. We propose Circle Attention, a novel spatial attention mechanism for telecom traffic forecasting, which directly models the area of effect of neighboring cell towers. Cell towers typically point in three different geographical directions, called sectors. Circle Attention models the relationships between sectors of neighboring cell towers by assigning a circle with learnable parameters to each sector, which are: the azimuth of the sector, the distance from the cell tower to the center of the circle, and the radius of the circle. To model the effects of neighboring time series, we compute attention weights based on the intersection of circles relative to their area. These attention weights serve as multiplicative gating parameters for the neighboring time series, allowing our model to focus on the most important time series when making predictions. The circle parameters are learned automatically through back-propagation, with the only signal available being the errors made in the traffic forecasting of each sector. To validate the effectiveness of our approach, we train a Transformer to forecast the number of attempted calls to sectors in the Copenhagen area, and show that Circle Attention outperforms the baseline methods of including either all or none of the neighboring time series. Furthermore, we perform an ablation study to investigate the importance of the three learnable parameters of the circles, and show that performance deteriorates if any of the parameters are kept fixed. Our method has practical implications for telecommunication operators, as it can provide more accurate and interpretable models for forecasting network traffic, allowing for better resource allocation and improved service provision

Currently, telco operators are experiencing decline in revenues and therefore exploring new business opportunities through innovative digital solutions directed towards industry and the public sector. Previous research on digitalization has focused on urban communities and smart cities and less on rural areas, where the digital divide is more frequent. The EU wants to revitalize local communities through enhanced connectivity-based solutions for businesses, farmers, institutions, and families. Eliminating the digital divide in rural communities is vital if we want to fully draw on the effects of digitalization with respect to increased efficiency and reduced climate and environmental impact. The Horizon Europe project COMMECT (Connecting Communities) is such a vehicle for the revitalization of rural industries in five countries in Europe, including Norway which hosts the Connected forestry Living Lab. This article proposes a method for assessing the social and economic impacts of connectivity-based solutions such as fifth generation (5G) mobile networks and the Internet of Things on the most important steps of the forestry value chain. The method proposes assessment of expected impacts early in the development process through trials of protocepts/proofs of concepts/minimum viable products in laboratory environments prior to more extensive field trials in the intended environments. The proposed method is novel and applicable to the assessment of the impacts of digitalization beyond the forestry industry.

The aquaculture industry has a goal of automating as much as possible to minimize cost and improve product quality. Cameras and environmental sensors are extensively used to monitor the fish farming sites, and generate huge amounts of data. 5G technology is seen as an enabler to further improve efficiency and digitize the fish farming industry. In this work, the combination of 5G, Device Edge, Cloud and Artificial Intelligence (AI) has been tested to evaluate the benefits and limitations of 5G technology in aquaculture. By emulating a typical Norwe-gian Atlantic salmon farm, remote monitoring, feeding decision support using AI for pellet detection, and 5G performance has been assessed. Peak uplink data rate is the most important key performance indicator, due to the large amount of data produced in the farm itself. To reduce the uplink requirements, a Device Edge has been deployed for running AI-driven pellet detection. Results show that operating full video coverage both underwater and for surveillance clearly exceeds the offered uplink data rate of a typical 5G base station operating in the C-band. Video compression can only be used to a mild extent, due to early deterioration of the pellet detection precision. Therefore, the use of a Device Edge to avoid uplink transmission of the video streams seems to be a better solution. Latency has not been critical in the scenario investigated, however introduction of remote control of cameras and feed provision might change this.

5G commercialization relies on validating the vertical use cases and selecting the ones creating values for both mobile operators and vertical stakeholders. To transfer the validations to commercialization more quickly, it is important to build a 5G platform not only with a similar scale and reliability as a commercial 5G but also capable of offering advanced beyond 5G services. In this paper, we propose an Experiment-as-a-Service (EaaS) framework which systematically offers four types of testing services, integration, functional, performance, and security testing. Then the iCORA (innovative, cloud-native, open, robust and automated) platform is presented to demonstrate how the EaaS framework could be realized in a large-scale 5G experimentation platform, which offers both diversity and flexibility for testing and experiments. A media use case is used to exemplify how the EaaS tailors the iCORA network to support various demands of vertical use cases and meet their KPIs.

The growing complexity and capacity demands for mobile networks necessitate innovative techniques for optimizing resource usage. Meanwhile, recent breakthroughs have brought Reinforcement Learning (RL) into the domain of continuous control of real-world systems. As a step towards RL-based network control, this paper introduces a new framework for benchmarking the performance of an RL agent in network environments simulated with ns-3. Within this framework, we demonstrate that an RL agent without domain-specific knowledge can learn how to efficiently adjust Radio Access Network (RAN) parameters to match offline optimization in static scenarios, while also adapting on the fly in dynamic scenarios, in order to improve the overall user experience. Our proposed framework may serve as a foundation for further work in developing workflows for designing RL-based RAN control algorithms.

5G network technology can offer major benefits to the Media & Entertainment (M&E) sector in addressing the challenges associated with production and delivery of mobile, IP-driven media services, such as smart media production, 4K/8k video streaming, and live virtual reality. The massive growth in demand for such advanced services has created the need for innovative experimentation facilities. These can support the M&E industry in continually enhancing user experience and creating efficiencies in how content is produced and delivered across mobile networks. Such support is the core goal of the H2020-funded 5GMediaHUB project, which aims to accelerate the testing and validation of innovative 5G-enabled media applications and Network Applications, through an open, integrated and fully featured experimentation facility. This will significantly reduce the service creation lifecycle and minimise time-to-market, thus enhancing the competitiveness of facility users. 5GMediaHUB will create an elastic, secure and trusted multi-tenant service execution and Network Application development environment, based on an open, cloud-based architecture and APIs. The project will integrate this testing and validation system with two well-established 5G testbeds at CTTC, Spain and Telenor, Norway. In this article, we first introduce the key features of the 5GMediaHUB experimentation facility. These features ensure effective operation of advanced media applications over 5G, and support developers in minimising application performance uncertainties prior to widescale commercial deployment. Next, the deployment and implementation of the 5GMediaHUB infrastructure is presented, before the deployment and validation of the project’s use cases are discussed. These use cases will explore the topics of Interactive Media, Content Creation and Media Distribution, all enabled through advanced 5G connectivity.

The diffusion of the Fifth-Generation (5G) of mobile radio networks will be the main driver in the digital transformation towards a new hyper-connected society. In order to satisfy the stringent demands of 5G-ready applications over the limited resources available at the edge, scaling mechanisms become crucial to guarantee the performance levels envisaged for 5G. Such mechanisms must be able to automatically perform according to the real-time user demands, the availability of computing resources and the state of Network Functions (NFs) and applications. In this context, this paper proposes a deep learning model, based on Artificial Neural Networks (ANNs), for the dynamic and automated orchestration of NFs. The novelty of this model is its independence from specific 5GNF implementations; this is due to the nature of the Key Performance Indicators (KPIs) used in this work, which are related to both execution environment (standard “black-box” KPIs) and standard 5G APIs (“white-box” KPIs). Results obtained on the orchestration of a Session Management Function (SMF) reach an accuracy of 97∼ 98% for the training and validation phases and above 95% for the deployed model, as well as higher overall accuracy by ∼ 5% and computational resource savings with respect to a threshold-based scheme.

Time series generation (TSG) studies have mainly focused on the use of Generative Adversarial Networks (GANs) combined with recurrent neural network (RNN) variants. However, the fundamental limitations and challenges of training GANs still remain. In addition, the RNN-family typically has difficulties with temporal consistency between distant timesteps. Motivated by the successes in the image generation (IMG) domain, we propose TimeVQVAE, the first work, to our knowledge, that uses vector quantization (VQ) techniques to address the TSG problem. Moreover, the priors of the discrete latent spaces are learned with bidirectional transformer models that can better capture global temporal consistency. We also propose VQ modeling in a time-frequency domain, separated into low-frequency (LF) and high-frequency (HF). This allows us to retain important characteristics of the time series and, in turn, generate new synthetic signals that are of better quality, with sharper changes in modularity, than its competing TSG methods. Our experimental evaluation is conducted on all datasets from the UCR archive, using well-established metrics in the IMG literature, such as Fréchet inception distance and inception scores.

The development of 5G and Beyond 5G (BSG) technologies relies on the availability of experimentation facil-ities that can evaluate and validate the performance of these technologies. It is of great interest and challenge to design, deploy and operate large-scale experimentation platforms to meet the high requirements of various vertical use cases for the 5G services. This paper describes an i-CORA platform that we build with multiple partners in Norway to support several EU-funded projects (5GMediaHUB, IMAGINE-B5G, FIDAL and COMMECT) and vertical use cases. The platform is cloud-native and consists of four parts: a multi-vendor end-to-end 5G network with three RAN sites serving general use cases, two mobile private networks (MPNs) and three Networks on Wheels (NOWs) serving dedicated verticals, and an open source platform composed of open source solutions. i-CORA offers both advanced standalone 5G services and value-added services (e.g., security and testing) to verticals in Public Protection and Disaster Relief (PPDR), media, eHealth, Industry 4.0, etc. In this paper, we address the challenges and lessons learned during the implementation and operation of the i-CORA platform.