Radio Link Failure (RLF) detection is essential for maintaining reliable connectivity in 5G networks. However, traditional centralized detection mechanisms often encounter scalability and latency constraints when managing large-scale, geographically distributed infrastructures. To address this challenge, we introduce a Lyapunov-driven federated learning framework that adaptively selects gNodeBs based on both data utility and historical participation. This approach leverages an LSTM-based local model to capture temporal patterns in link performance, thereby enhancing RLF detection. Extensive evaluations on a real-world 5G dataset demonstrate that the proposed method achieves superior performance compared to baseline approaches when detecting rare failure events. By simultaneously prioritizing performance and fairness, this framework offers a scalable solution suited to diverse and dynamic 5G environments.

Unmanned Aerial Vehicles (UAVs) are expanding beyond military use into sectors such as logistics, communication, and transportation. However, their dependence on high-power batteries limits their range, and while fuel cells provide longer flight times, they can reduce speed and acceleration due to safety concerns. Consequently, managing UAV power consumption has become a critical challenge, directly affecting flight duration and operational performance. Therefore, accurately predicting power consumption is important for enabling efficient UAV mission planning. Thus, in this paper, we propose a fine-tuning strategy called Large Language Model for Power Forecasting (LLM4PF) that employs a Generative Pretrained Transformer (GPT-2) model to reduce computational costs without sacrificing accuracy. LLM4PF predicts power consumption based on various UAV operational data including speed and altitude among others. Furthermore, we evaluate its performance in low-data scenarios through few-shot learning with 5% and 10% data subsets. Additionally, we compare LLM4PF to transformer-based models using a public dataset, demonstrating its effectiveness and efficiency.

Radio Link Failure (RLF) detection in Radio Access Networks (RANs) is crucial for ensuring seamless communication in 5G networks. Nonetheless, current approaches based on traditional Machine Learning (ML) algorithms fail to find a trade-off between accuracy and computational complexity. Thus, in this paper, we explore advanced and computationally efficient types of transformers, such as Linformer and Performer. These models significantly reduce computational complexity while maintaining strong performance by leveraging different attention mechanisms. Extensive evaluations using a realistic dataset under various percentages of link failures show that Linformer provides a favourable balance between accuracy and training time.

Internet of Vehicles (IoV) frameworks integrate smart vehicles, roads, network infrastructures, and users into one system, enhancing environmental awareness, increasing efficiency, and reducing accidents. Although IoV is widely adopted, it has resulted in an increase in global energy demand, necessitating more robust and reliable energy solutions. In order to meet this growing demand, both traditional and distributed energy generation technologies have been developed, particularly renewable energy sources (RES). In order to ensure seamless operation of smart vehicles and related infrastructure, integrating renewable energy into existing grid infrastructure is essential. This enables stable and efficient power supply to IoV systems, especially during time periods of high demand. As part of this shift, it is important to accurately forecast the energy generation load of individual prosumers—entities that both produce and consume energy—because of their intermittent and dynamic nature. Therefore, we propose Block-FeST, a blockchain-based Federated Learning (FL) framework designed to predict the energy generation patterns of RES prosumers while preserving their private and sensitive data. Within this Block-FeST framework, a Sparse Transformer model is used to forecast energy generation among prosumer clients. Additionally, blockchain technology is integrated into the Block-FeST framework to enable distributed aggregation and securely validate and record the local parameters shared by clients. The results indicate that Block-FeST is superior to the second-best baseline method, with improvements of 20.4% in the mean square error (MSE), 13.7% in mean absolute error (MAE) and 19.3% in root mean square error (RMSE) for a long sequence length of 128.

Internet of Things (IoT) applications generate tremendous amounts of data streams which are characterized by varying Quality of Service (QoS) indicators. These indicators need to be accurately estimated in order to appropriately schedule the computational and communication resources of the access and Edge networks. Nonetheless, such types of IoT data may be produced at irregular time instances, while suffering from varying network conditions and from the mobility patterns of the edge devices. At the same time, the multipurpose nature of IoT networks may facilitate the co-existence of diverse applications, which however may need to be analyzed separately for confidentiality reasons. Hence, in this paper, we aim to forecast time series data of key QoS metrics, such as throughput, delay, packet delivery and loss ratio, under different network configuration settings. Additionally, to secure data ownership while performing the QoS forecasting, we propose the FeDerated Temporal Sparse Transformer (FeD-TST) framework, which allows local clients to train their local models with their own QoS dataset for each network configuration; subsequently, an associated global model can be updated through the aggregation of the local models. In particular, three IoT applications are deployed in a real testbed under eight different network configurations with varying parameters including the mobility of the gateways, the transmission power and the channel frequency. The results obtained indicate that our proposed approach is more accurate than the identified state-of-the-art solutions.

Internet of Things (IoT) devices generate a tremendous amount of time series data that is extremely dynamic, heterogeneous and time dependent. Such types of data introduce significant challenges for the real-time prediction of QoS metrics of IoT applications with different traffic characteristics. To this end, in this paper, we propose a temporal transformer model and a unified system to predict several QoS metrics of heterogeneous IoT applications when they communicate with the Edge of the network. The transformer model also leverages an attention module to provide a solution for both short-term and long-term sequence prediction of QoS metrics that allows to better extract any time dependencies. In particular, in our framework, we firstly generate a set of datasets containing real-time traffic information of five different IoT applications such as Heating, Ventilation, and Air Conditioning (HVAC), lighting, Voice over Internet Protocol (VoIP), surveillance and emergency response using the 802.15.4 access technology and the RPL routing protocol. Following, we perform the data cleaning, downsampling and pre-processing of the datasets and we construct the QoS datasets, which include four QoS metrics, namely throughput, packet delivery ratio, packet loss ratio and latency. Finally, we evaluate the transformer model through extensive experimentation using both short-term and long-term dependencies and we show that our model can guarantee a robust performance and accurate QoS prediction.