Harnessing the potential of smart grid networks relies on the efficient and secure communication between distributed energy resources (DERs) and the energy management system (EMS), particularly under variable channel conditions and adversarial interference. This interference is further exacerbated with the advent of artificial intelligence (AI)-driven adversarial devices capable of adaptively disrupting communication in real time. To address these challenges, in this study, we formulate the distributed channel access problem, incorporating dynamic channels and intelligent adversarial interference as a partially observable Markov game (POMG). Specifically, we propose a distributed framework based on a centralized training and distributed execution (CTDE) multi-agent reinforcement learning (MARL), enabling DERs to autonomously adapt to dynamic channels and mitigate intelligent interference using only local observations, i.e., without direct information sharing among DERs. Our simulation results indicate that, by integrating an advanced policy evaluation technique and a tailored utility maximization strategy, DERs can collaboratively optimize their transmission decisions, improving the network's aggregate packet success rate (APSR) and resilience. Additionally, the proposed framework outperforms existing methods, ensuring robust and scalable communication in smart grids under diverse conditions.

This paper addresses the challenges faced by 6G-enabled vehicular networks (V-Nets), including increasing road traffic, ultra-reliable and low latency communication, high data rates, and energy efficiency. The intelligent reflecting surface (IRS) is proposed as a solution to configure the propagation channel in a smart radio environment by adjusting phase shifts. However, designing IRS-assisted V-Nets that achieve ultra-reliability in dynamic and noisy communication is challenging due to the passive nature of the IRS and the limitations of deep reinforcement learning (DRL) methods. To overcome these challenges, this paper presents a computer vision (CV) enabled IRS framework for V-Nets, which combines a convolutional neural network and CV techniques. The framework utilizes real-time visual information to estimate and configure optimal beamforming for IRS-assisted V-Nets. Adapting to real-time network dynamics and intelligently guiding signals, the CV-IRS framework improves prediction accuracy to 95%, an achievable maximum rate of 11.2 bps/Hz with 100 IRS elements, and resource allocation efficiency of 88% with 10 vehicles. The simulation results demonstrate the superiority of the CV-IRS framework over benchmark schemes, making it a promising approach for the efficient configuration of IRS-assisted 6G V-Nets.

Traditional human-designed medium access control (MAC) protocols cannot tackle the heterogeneous requirements of the future 6G wireless networks. Reinforcement learning (RL) algorithms have been proposed, in which base stations (BSs) and user equipment's (UEs) act as agents to automatically learn the MAC protocols to satisfy the stringent quality of service (QoS) requirements of 6G networks. However, existing RL techniques result in a generalization issue where agents fail to identify and explore useful information in a sparse wireless environment. To tackle this challenge, we propose a cooperative multi-agent exploration (CMAE) framework in which the network state space is projected into a low-dimensional space instead of learning a policy in a high-dimensional space. Consequently, the agents start exploring from low-dimensional state space to high-dimensional space to learn the abstracted information from the wireless environment. In the proposed framework, the nodes and BSs collaborate to explore the under-explored wireless network states to jointly learn the channel access and signalling policy. Simulation results show that the proposed CMAE framework outperforms traditional baseline schemes in terms of good put and collision rate and has better generalization capabilities.

Tactical wireless sensor networks (T-WSNs) are used in critical data-gathering military operations, such as battlefield surveillance, combat monitoring, and intrusion detection. These networks have unique challenges, such as jamming attacks, which are not normally encountered in traditional WSNs. Jamming attacks on the networks’ links disrupt data communication and make packet routing in T-WSNs a difficult task. Consequently, T-WSN routing aims to find the most reliable routes, while meeting the stringent delay and energy requirements. To this end, we propose a distributed multi-agent deep reinforcement learning (MADRL)-based routing solution for multi-sink tactical mobile sensor networks to overcome link layer jamming attacks. Our proposed routing scheme captures the hop count to the nearest sink, the one-hop delay, the next hop’s packet loss rate (PLR), and the energy cost of packet forwarding in the action reward estimation. Furthermore, the proposed scheme outperforms benchmark algorithms in terms of the packet delivery ratio (PDR), packet delivery time, and energy efficiency.

The performance of wireless networks is constrained by the dynamic and random nature of the wireless channels. Intelligent reflecting surface (IRS) is a promising approach that can smartly reconfigure wireless propagation environment to increase the spectral efficiency in 6G networks. However, IRS deployment optimization in a complex and random 6G environment remains a limiting factor in improving the performance. To address the issue, we propose a deep reinforcement learning (DRL) network empowered by a generative adversarial network (GAN) to jointly optimize the IRS placement and reflecting beamforming matrix of IRS as well as the transmit beamforming at the base station (BS) in an IRS-assisted wireless network. Simulation results show that the proposed technique outperforms the benchmark scheme in terms of achievable rate and signal-to-noise ratio (SNR) by learning the optimal IRS locations in an IRS-aided wireless network.

Recent advances in electronic devices and communication infrastructure have revolutionized the traditional healthcare system into a smart healthcare system by using internet of medical things (IoMT) devices. However, due to the centralized training approach of artificial intelligence (AI), mobile and wearable IoMT devices raise privacy issues concerning the information communicated between hospitals and end-users. The information conveyed by the IoMT devices is highly confidential and can be exposed to adversaries. In this regard, federated learning (FL), a distributive AI paradigm, has opened up new opportunities for privacy preservation in IoMT without accessing the confidential data of the participants. Further, FL provides privacy to end-users as only gradients are shared during training. For these specific properties of FL, in this paper, we present privacy-related issues in IoMT. Afterwards, we present the role of FL in IoMT networks for privacy preservation and introduce some advanced FL architectures by incorporating deep reinforcement learning (DRL), digital twin, and generative adversarial networks (GANs) for detecting privacy threats. Moreover, we present some practical opportunities for FL in IoMT. In the end, we conclude this survey by discussing open research issues and challenges while using FL in future smart healthcare systems.

Next-generation wireless communication and networking technologies, such as sixth-generation (6G) networks and software-defined Internet of Things (SDIoT), make cyber-physical systems (CPSs) more vulnerable to cyberattacks. In such massively connected CPSs, an intruder can trigger a cyberattack in the form of false data injection, which can lead to system instability. To address this issue, we propose a graphics-processing-unit-enabled adaptive robust state estimator. It comprises a deep learning algorithm, long short-term memory, and a nonlinear extended Kalman filter, and is called LSTMKF. Through an SDIoT controller, it provides an online parametric state estimate. The reliability is improved by performing two levels of online parametric state estimation for secure communication and load management. The CPS under study is a 6G and SDIoT-enabled smart grid, which is tested on IEEE 14, 30, and 118 bus systems. Compared to existing techniques, the proposed algorithm is able to estimate the state variables of the system even during or after a cyberattack, with lower time complexity and high accuracy.

The traditional Internet architecture relies on the best-effort principle, which is not suitable for critical industrial Internet of Things (IIoT) applications such as healthcare systems with stringent quality-of-service (QoS) requirements. In this article, a software-defined network (SDN) based on an analytical parallel routing framework is proposed by using the massive processing power of a graphics processing unit (GPU) for dynamically optimizing multiconstrained QoS parameters in the IIoT. The framework considers three types of QoS applications for smart healthcare traffic: loss-sensitive, delay-sensitive, and jitter-sensitive. A QoS-enabled routing optimization problem is formulated as a max-flow min-cost problem, while a greedy heuristic that dispatches the path calculation task concurrently to the GPU for calculating optimal forwarding paths considering the QoS requirement of each flow is proposed. The results show that the proposed scheme efficiently utilizes the limited bandwidth cost in terms of energy and bandwidth while satisfying the QoS requirement of each flow with maximizing the network resources for future IIoT traffic flows. Comparative analysis of simulation results with shortest path delay, Lagrangian relaxation-based aggregated cost, and Sway schemes indicate a reduced violation in the service-level agreement by 17%, 19%, and 4%, respectively, by using the AttMpls topology, while it is 48%, 44%, and 7% when the Goodnet topology is used. Moreover, SEQOS is seen to be energy efficient and eight times faster than the benchmark algorithms in large IIoT networks.

The Cognitive Internet of Vehicles (CIoV) is an intelligent network that embeds the cognitive mechanism in the Internet of Vehicles (IoV) to sense the environment and observe the network states to learn the optimal policies adaptively. However, one of the key challenges in CIoV systems is to design a smart agent that can smartly schedule the packet transmission for ultra-reliable low latency communication (URLLC) under extreme random and noisy network conditions. We propose a software defined network (SDN) based scheduling algorithm that leverages generative adversarial network (GAN) based deep distributional Q-network (GAN-DDQN) for learning the action-value distribution for intelligent transmission scheduling. A reward-clipping technique is proposed for stabilizing the training of GAN-DDQN against the effect of broadly spanning utility values. The extensive simulation results verify that GAN-Scheduling achieves higher spectral efficiency (SE), service level agreement (SLA), system throughput, transmission packet rate with lower transmission delay, and power consumption compared to the existing reinforcement learning algorithms.

Internet of things (IoT) aims to connect billions of devices to the Internet. These devices generate an enormous amount of traffic. A Wireless Body Area Network (WBAN) aggregates the data generated from the body sensors deployed on the human body or inside it. Due to resource constraints of the deployed sensors, aggregated data is forwarded to a server at an edge node or a cloud for further processing and decision making. With the wide acceptance of IoT deployments, WBANs can also be integrated with IoT to provide a clear contextual environment for smart healthcare applications' users. Software Defined Networks (SDN) is a new networking paradigm providing centralized control and programmability of the networks. These features of SDN open new vistas for WBAN and IoT applications. In this paper, a personal digital assistant (PDA) is configured as an SDN switch, and we emulate the SDN functionality with Mininet to verify that this approach can be used in WBANs with minimum deployment complexity and network overhead, while efficiently manage the connected devices and the connections between them.