Georgios Ellinas

In the coming years, unmanned aerial vehicles (UAVs) are expected to find applications in various new mobility concepts. However, current practical trajectory planning solutions for these vehicles in densely populated environments either inadequately address safety concerns, under-utilize available airspace, or rely on overly optimistic assumptions. To address these challenges, we introduce a novel Unmanned Traffic Management System (UTMS), which first partitions the airspace into cubic regions of equal size and employs a reservation mechanism to plan inter-cube paths from source to destination. In this way the proposed UTMS reservation mechanism can accommodate different quality-of-service (QoS) levels and admit UAVs within cubes up to a (tunable) maximum allowed density. Secondly, within each cube, an intra-cube planner robustly determines the trajectory of each UAV, ensuring a guaranteed safety level under realistic system state uncertainty. Extensive simulation results demonstrate that the proposed UTMS effectively manages the trade-off among the tunable UAV density and safety, thereby mitigating the design limitations observed in current state-of-the-art approaches.

Within the trajectories of design-based research (DBR) and multimodal theory, this study captures the iterative processes undertaken to design, develop, and deploy game-based Augmented Reality (AR) applications for learning, exploration, and collaboration within higher education settings. Three game-based scenarios were designed set in a prehistoric neolithic site, Choirokoitia, in the Mediterranean region. The study investigated the pedagogical and technological affordances, complexities, and contradictions that our interdisciplinary team experienced during these iterative processes. Using a multimodal perspective and drawing on Klopfer and Squire’s (2008) framework, we delineated six phases, from aligning the AR scenario design with European project requirements and creating multimodal contexts, to identifying future research directions. The results demonstrate the intricacies and iterative processes utilised in harnessing multiple software and hardware resources for the conception and realisation of AR embedded within culturally and historically rooted contexts that enact affordances for exploration and collaboration. These endeavours were underpinned by pedagogical imperatives that focus on task-oriented activities.

This work extends the distributed Nash equilibrium seeking algorithm to address multi-agent games under time-varying, signed, and undirected communication networks. Initially, based on uniform complete observability for linear time-varying systems, a new analytical framework is developed to determine the necessary and sufficient condition for achieving globally uniformly exponential stability of the error dynamics of the distributed estimator. It is shown that exponential stability is realized if and only if the joint $(\epsilon,T)$-connectivity and the negative-link assumption of time-varying signed graphs are satisfied. Subsequently, it is demonstrated that under some mild conditions, the seeking algorithm ensures the global uniform exponential convergence of players' strategies towards the Nash equilibrium.

The implementation of a set of projects within an organization can be scheduled through a project scheduling procedure, that is based on the organizational resources available, such as the number of teams and team members, and the tasks that each team must perform. Nevertheless, hiring new personnel (staff hiring) and rearranging the members among the existing teams (staff reassignment) can lead to better scheduling solutions. This can potentially have a large financial impact for the organization, as it will reduce the projects’ implementation cost and will bring products earlier to market. To this end, this work presents different approaches to solve the resource-constrained project scheduling problem considering the reassignment of existing personnel as well as the hiring of new personnel. This is achieved by modifying a previously presented integer linear program (ILP) formulation to solve the aforementioned problem and considering different objectives and scenarios for staff hiring and reassignment. From the results obtained, it is evident that by using the developed ILPs, an organization can schedule its projects in a more time- and cost-efficient manner through the newly proposed recruitment and staff reassignment processes, compared to the benchmark approach that considers only existing personnel and team compositions.

With the increasing technological advancements and deployment of unmanned aerial vehicles (UAVs) in different applications (e.g., delivery services, surveillance, critical infrastructure monitoring, military operations, and search-and-rescue operations), a robust, lightweight, and accurate system for UAV state identification and trajectory prediction is becoming a mandate. This work introduces an onboard real-time multi-task learning framework that combines Transformer neural networks and reservoir computing (RC) architectures to enhance outdoor UAV operations. The proposed system employs the Transformer-based architecture to capture the temporal dependencies in sequential data for long-term horizons, while RC-based networks are utilized to ensure robust and real-time performance. Specifically, custom multi-task models are implemented and fine-tuned to collect multi-modal sensor measurements, aiming to enhance the two-fold objective of simultaneous UAV state identification and trajectory prediction through shared feature learning. A dataset comprised of measurements collected from real-world UAV operations (in 3D space) under various conditions is employed to train and evaluate the proposed system. Furthermore, a prototype onboard UAV system is implemented and tested in real-world field experiments. Extensive evaluations of the experimental results demonstrate that the onboard framework achieves accuracy close to the ground truth in both state identification and trajectory prediction, consequently showcasing its potential for practical applications.

In this work, the network coding (NC) technique is used in combination with routing and spectrum allocation (RSA) to establish confidential connections in elastic optical networks and protect these connections against eavesdropping attacks. Utilizing NC, the signals of the confidential connections are encrypted using other signals at different nodes in their path while transmitted through the network, preventing an eavesdropper tapping a link to gain access to confidential information. A novel mixed integer linear program (MILP) is proposed to enable encrypted transmission of confidential connections, in combination with a mapping function that maximizes the security level provided for the largest possible set of confidential connections. Further, a heuristic approach is presented for the combined routing, spectrum, and network coding assignment (RSNCA) problem, along with a metaheuristic, for larger-sized networks. The proposed approaches are examined in terms of confidentiality, average number of encryption operations (EOs), spectrum utilization, and running times. Performance results demonstrate the applicability of the MILP formulations in providing optimal lightpath establishments that maximize physical layer security, with the proposed MILP-RSNCA approach achieving the best results in terms of the level of security and the average number of EOs provided, albeit utilizing more spectrum slots. Further, the metaheuristic exhibits results comparable to the MILP in terms of providing security, and can be also applied to large-size networks. In some scenarios, it can even provide the highest available level of security i.e., a minimum of 7 EOs per link per demand for more than 50% of the confidential demands.

This paper draws from over eight years of experience in designing and developing augmented reality (AR) applications focused on culture, history, second and foreign language learning, multilingualism, and other salient values. Guided by Design-based research (DBR) (Amiel & Reeves, 2008), the study captures the multifaceted, complex, and collaborative processes involved in designing, developing, and deploying AR applications for the DiMPAH and DIMPE projects. An interdisciplinary team of researchers, a software developer, and experienced instructors collaboratively navigate through the complex iterative processes. The study demonstrates that the design and development processes are galvanized by problem analysis, pedagogical and technological values, iterative testing and improvement, and reflection principles. The nature of each European project determines the design and development of the AR applications, as well as the task-driven activities, learning experiences, and intended pedagogical values that are shaped by institutional requirements and instructors’ and students’ teaching and learning needs.

The wide adoption and use of unmanned aerial vehicles (UAVs) has created not only opportunities but also threats to the security of sensitive areas. Thus, effective and efficient counter-drone systems are required to protect these areas. This work tackles this issue by developing cooperative multi-agent jamming techniques using reinforcement learning (RL) to counter the operation of one or multiple rogue drones flying over a sensitive area. The aim of the proposed RL approach is to optimize the joint mobility and power control actions of the pursuer UAVs in order to maximize the received jamming power at the rogue drones aiming at disrupting communication links and sensing circuitry, while at the same time keeping the interference to surrounding pursuer agents below a predefined threshold. The effectiveness of the proposed approach in terms of scalability, learning speed, and agents’ final joint performance is demonstrated through extensive simulation experiments for various agent and target configurations.

A multi-period planning framework is proposed that exploits multi-step ahead traffic predictions to address service overprovisioning and improve adaptability to traffic changes, while ensuring the necessary quality-of-service (QoS) levels. An encoder-decoder deep learning model is initially leveraged for multi-step ahead prediction by analyzing real-traffic traces. This information is then exploited by multi-period planning heuristics to efficiently utilize available network resources while minimizing undesired service disruptions (caused due to lightpath re-allocations), with these heuristics outperforming a single-step ahead prediction approach.