Energy efficiency is a huge opportunity for both the developed and the developing world, and ICT will be the key enabler towards realising this challenge, in a huge variety of ways across the full range of industries. In the telecommunications space in particular, power consumption and the resulting energy-related pollution are becoming major operational and economical concerns. The exponential increases in network traffic and the number of connected devices both make energy efficiency an increasingly important concern for the mobile networks of the (near) future. More specifically, as 5G is being deployed at a time when energy efficiency appears as a significant matter for the network ability to take into account and to serve societal and environmental issues, this can play a major role in helping industries to achieve sustainability goals. Within this scope, energy efficiency has recently gained its own role as a performance measure and design constraint for 5G communication networks and this has identified new challenges for the future. In particular, the inclusion of AI/ML techniques will further enhance 5G’s capabilities to achieve lower power consumption and, most importantly, dynamic adaption of the network elements to any sort of energy requirements, to ensure effective functioning.

Machine learning-based semantic communication is a promising enabler for future-generation wireless network systems such as 6G networks. In practice, effective semantic communication requires online training for unknown content. In highly mobile vehicular networks, however, reliable, and efficient model training becomes significantly challenging. The existing distributed learning approaches are also unable to effectively operate in highly dynamic vehicular semantic communication networks. To address these challenges, we propose a novel mobility-aware split-federated with transfer learning (MSFTL) framework based on vehicle task offloading scenarios in this paper. To enable adaptation to the complex vehicle semantic communication, the proposed framework divides the training of the model into four parts and uses the proposed new splitfederated learning. Furthermore, to improve training efficiency, model accuracy, and the ability to adapt in highly mobile environments, we also present a new transfer learning approach integrated into the proposed framework. Particularly, we propose a high-mobility training resource optimisation mechanism based on a Stackelberg game for MSFTL to further reduce training costs and adapt vehicle mobility scenarios. We also investigate the performance of the proposed schemes through extensive simulations. The results validate the proposed approach and indicate its superiority compared to the conventional learning frameworks for semantic communication in vehicular networks.

Modern applications involve green communication technologies motivating well optimization in the power-limited regime. In comparison with most of the existing related work that assumes perfect channel state information (CSI) is always available, which is unfortunately not true in reality, this paper focuses on an optimal energy-efficient solution for resource allocation in multiuser orthogonal frequency division multiple access networks in the presence of imperfect CSI and data outage conditions. In particular, in view that wireless channel conditions, circuit power consumptions, and users' quality-of-service (QoS) requirements are heterogeneous in nature, we enable attractive tuning options by letting energy efficiency optimization objective to assign weights to each allocation link. In addition, we interpret the effects of data outage due to imperfect CSI using a profound insight on the monotonicity of noncentral chi-squared inverse distribution function, which reveals that our design complies with expected physics and mechanics of conventional energy efficiency approach and that it can be successfully degenerated to the energy-efficiency model with perfect CSI. Furthermore, we formulate a mixed combinatorial problem toward maximizing the energy efficiency subject to a minimum QoS requirement, channel interference, and transmitting power constraints. The problem is transformed into an equivalent quasi-concave problem with respect to power, and concave problem with respect to the subcarrier indexing coefficients using the concept of subcarrier time sharing. We optimize through a simple and versatile methodology, which uses standard-Lagrangian optimization technique to obtain joint dynamic subcarrier and adaptive power allocations by means of final formulas. We also examine key properties of the introduced optimal solution in terms of implementation convergence and complexity, level of optimality, and impact of imperfect CSI coefficients and circuit power on network performance. The simulation results demonstrate the effectiveness of our allocation scheme for achieving higher energy efficiency performance with the guaranteed QoS support and lower complexity than the existing approaches especially when perfect CSI is not available.

Semantic communication based on machine learning (ML) techniques emerged as a new transmission paradigm that can significantly improve spectrum efficiency. It looks promising for improving the task of offloading quality of service (QoS) for autonomous driving networks (ADNs) where autonomous vehicles require a significant amount of communication with the vehicle edge clouds (VECs). However, in practical ADNs, updating the ML-based semantic communication coder model is affected by various unique factors such as mobility and privacy considerations. Therefore, in ADNs, the conventional ML frameworks are not directly applicable to updating semantic communication coders. In this article, we discuss the unique challenges faced by updating the semantic communication coder in ADNs, and review the existing ML frameworks. To address these challenges, we further propose a privacy-preserving personalized federated learning (3PFL) framework for updating the semantic communication coder in ADNs. Simulation results confirm the effectiveness of 3PFL for this process.

The 5G communication network will underpin a vast number of new and emerging services, paving the way for unprecedented performance and capabilities in mobile networks. In this setting, the Internet of Things (IoT) will proliferate, and IoT devices will be included in many 5G application contexts, including the Smart Grid. Even though 5G technology has been designed by taking security into account, design provisions may be undermined by software-rooted vulnerabilities in IoT devices that allow threat actors to compromise the devices, demote confidentiality, integrity and availability, and even pose risks for the operation of the power grid critical infrastructures. In this paper, we assess the current state of the vulnerabilities in IoT software utilized in smart grid applications from a source code point of view. To that end, we identified and analyzed open-source software that is used in the power grid and the IoT domain that varies in characteristics and functionality, ranging from operating systems to communication protocols, allowing us to obtain a more complete view of the vulnerability landscape. The results of this study can be used in the domain of software development, to enhance the security of produced software, as well as in the domain of automated software testing, targeting improvements to vulnerability detection mechanisms, especially with a focus on the reduction of false positives.

This article investigates a control system within the context of sixth-generation wireless networks. The remote control performance optimization confronts the technical challenges that arise from the intricate interactions between communication and control sub-systems, asking for a co-design. Considering the system dynamics, we formulate the sequential co-design decision-makings of communication and control over a discrete time horizon as a Markov decision process, for which a practical offline learning framework is proposed. Our proposed framework integrates large language models into the elements of reinforcement learning. We present a case study on the age of semantics-aware communication and control co-design to showcase the potential of our proposed learning framework. Furthermore, we discuss the open issues remaining to make our offline learning framework feasible for real-world implementations and highlight the research directions for future explorations.