Deep reinforcement learning (DRL) is a powerful method for Dynamic Power Allocation (DPA) due to its adaptability to changing wireless environments. Crafting proficient DRL agents necessitates a well-structured problem representation, learning scheme, and agent interaction framework. Additionally, the environment, as the second pivotal component in DRL, must mirror realistic channel dynamics faithfully. This research tackles these often-neglected aspects. We introduce Het-DRL, a heterogeneous multi-agent DRL framework with a competitive interaction paradigm for DPA, moving away from conventional collaborative and centralized learning schemes. The independent learning scheme eliminates information exchange among agents, empowering autonomous decision-making. Moreover, we present our CELLUCARLA wireless simulator, an extension of the CARLA (Car Learning to Act) simulator with a cellular network layer, enabling realistic urban mobility modeling. We leverage CELLUCARLA to evaluate the proposed Het-DRL, highlighting its effectiveness for DPA in Orthogonal Frequency-Division Multiplexing (OFDM) systems, with up to 153% sum rate relative to the Weighted Minimum Mean Square Error (WMMSE) algorithm.

Dynamic Power Allocation (DPA) in wireless networks has been tackled utilizing model-driven optimization, reinforcement learning, and deep learning methods. However, these methods still face some limitations, including dependence on task-specific designs and limited adaptability to diverse network configurations. In this paper, we optimize three adaptation variants of a pre-trained 3.82-billion-parameter Large Language Model (LLM) for DPA. These strategies include LLM with Prompt Engineering (LLM-PE), which employs DPA-specific prompt engineering with detailed network and channel information; LLM with Retrieval Augmented Generation (LLM-RAG), which enhances prompts with retrieved optimal power allocation examples; and LLM with Fine-tuning (LLM-FT), which fine-tunes the model on a small DPA-specific dataset. Our results show that LLM-FT excels in adaptability across various network configurations and resilience to channel estimation errors, maintaining a sum rate performance above 99% compared to the optimal water-filling algorithm, using only 5000 fine-tuning examples.

Despite recent advancements in millimeter-wave (mm Wave) communication systems, the practical validation of research ideas in this field is still challenging. This is mostly due to the simplifying assumptions in their system modeling and overlooking hardware limitations in the simulation environment. In this paper, we propose a high-speed and programmable mm Wave testbed with multi-user (MU) multiple-input multiple-output (MIMO) beamforming capability. The baseband signal processing is preformed fully on a general-purpose processor. The radio frequency (RF) frontend holds a hybrid MIMO architecture, where each RF chain is connected to one phased-array antenna. The testbed allows to plug and play different physical layer signal processing algorithms or radio resource allocation techniques, including machine learning models, to assess their performance under real-life conditions.

Edge computing is a sustainable network paradigm that supports resource constrained Industrial Internet of Things (IIoT) devices in performing data-driven tasks. The distributed architecture of computing resources, combined with the heterogeneous and sporadic characteristics of IIoT tasks, presents a substantial challenge to achieving energy-efficient and low-latency task execution. In this paper, we demonstrate that modeling the Age of Information (AoI) as an exponential function can significantly enhance the performance of task scheduling and execution in edge computing systems. Furthermore, we introduce a low over-head Multi-Agent Deep Reinforcement Learning (DRL) algorithm with No Observation (MA-NO) for distributed edge computing resource allocation, where agents (IIoT devices) operate independently without communicating with each other. However, to prevent the divergence of the agents and ensure a Pareto-optimal decision-making, we adopt a common reward that is shared among all agents. That is, all agents receive the same reward, based on their collective performance. Simulation results show that our proposed AoI model reduces the average execution latency of the tasks by 35.2% compared to a linear AoI ones. Furthermore, our proposed MA-NO algorithm reduces the total energy consumption of the IIoT devices by 59.9% and 63.4% compared to the single-agent DRL algorithm and multi-agent DRL algorithm with partial observations, respectively.

We propose an extremely energy efficient and reconfigurable accelerator for performing Deep Neural Network (DNN) inferences on a Field-Programmable Gate Array (FPGA). Our design allows on the fly reconfigurability of the model to adapt it to any arbitrary DNN inference task and perform it with an extremely low latency, on the scale of tens of micro seconds. Our design can be adopted by resource constrained Industrial Internet of Things (IIoT) devices (e.g., mobile robots) to fulfill different DNN inference tasks. By parallelization of the matrix multiplications in each layer of the DNN and adopting an adjustable sliding window at the synthesize stage, our accelerator strikes a balance between the computing latency and the scarce resource utilization on the FPGA. It also uses a low-overhead control scheme for the accelerator’s internal interactions to update the DNN model on-the-fly. The results show that for a DNN model with 64 input features, 9 hidden layers (each with 64 neurons), and 64 output variables, our proposed inference accelerator speeds up the inference by 28.8 times and decreases the energy consumption by 31 % compared to an AMD Deep Learning Processing Unit.

With the advent of the multimodal immersive communication system, people can interact with each other using multiple devices for sensing, communication and/or application level control either onsite or remotely. As a breakthrough concept, a distributed sensing, computing, communications, and control (DS3C) fabric is introduced in this paper for provisioning 6G services in multi-tenant environments in a unified manner. The DS3C fabric can be further enhanced by natively incorporating intelligent algorithms for network automation and managing networking, computing, and sensing resources efficiently to serve vertical use cases with extreme and/or conflicting requirements. As such, the paper proposes a novel end-to-end 6G system architecture with enhanced intelligence spanning across different network, computing, and business domains, identifies vertical use cases and presents an overview of the relevant standardisation and pre-standardisation landscape.

Physical-Layer Network Coding (PNC) is an effective technique to improve the throughput and latency in wireless networks. However, there are two major challenges for PNC, especially when using higher order modulations: 1) phase synchronization and power control at the paired User Equipments (UEs); and 2) the ambiguity removal of the PNC mapping at the relay node. To address these challenges, in this paper, we apply power control at transmitting UEs and exploit Reconfigurable Intelligent Surfaces (RISs) to synchronize the phase of the transmitted signals and ensure that they arrive at the relay with the same power and phase rotation. Then, we employ modular addition for an unambiguous PNC mapping for M-ary Quadrature Amplitude Modulations (M-QAM). We evaluate the performance of the system in the framework of Orthogonal Frequency Division Multiplexing (OFDM)-PNC for different RIS sizes and modulation orders. Furthermore, we study the sensitivity of PNC systems for Channel Estimation Error (CEE). The results reveal that 1) PNC systems show quite higher sensitivity to CEE compared with RIS-assisted one-way relay channel systems; 2) when the CEE is low, RIS can considerably enhance the Signal-to-Noise Ratio (SNR) of the PNC system, e.g., for a Bit Error Rate (BER) of 10−3(without channel coding), increasing the RIS size from one to 256 elements in 28 GHz band leads to 200% improvement in SNR.

Integrated Access and Backhaul (IAB) has been recently proposed by 3GPP to enable network operators to deploy fifth generation (5G) mobile networks with reduced costs. In this paper, we propose to use IAB to build a dynamic wireless backhaul network capable to provide additional capacity to those Base Stations (BS) experiencing congestion momentarily. As the mobile traffic demand varies across time and space, and the number of slice combinations deployed in a BS can be prohibitively high, we propose to use Deep Reinforcement Learning (DRL) to select, from a set of candidate BSs, the one that can provide backhaul capacity for each of the slices deployed in a congested BS. Our results show that a Double Deep Q-Network (DDQN) agent using a fully connected neural network and the Rectified Linear Unit (ReLU) activation function with only one hidden layer is capable to perform the BS selection task successfully, without any failure during the test phase, after being trained for around 20 episodes.

Mobile user authentication is the primary means of verifying the claimed identity of a user before granting access to resources on a mobile device. Common user authentication methods include passwords and biometrics. Despite the fact that passwords have been the most popular user authentication method for several decades, recent research suggests that they are no longer secure or convenient for mobile users due to several limitations that compromise both device security and usability. Biometric-based user authentication, on the other hand, is gaining popularity because it appears to strike a balance between security and usability. Such methods rely on human physical traits (physiological biometrics) or user involuntary actions (behavioral biometrics) for authentication. Risk-based user authentication using behavioral biometrics is particularly promising for mobile user authentication enhancing mobile authentication security while maintaining usability. In this context, we present an overview of mobile user authentication and discuss risk-based user authentication for mobile devices as a suitable approach to deal with the security vs. usability challenge. Afterwards, we test and evaluate a set of outlier detection algorithms for risk estimation in order to identify the most suitable ones for risk-based user authentication on mobile devices in terms of their accuracy and efficiency.

Physical-layer Network Coding (PNC) has great potential to improve the throughput and latency in wireless networks. However, there are two main challenges in PNC systems that do not exist in the conventional Point-to-Point$(\mathrm{P}2\mathrm{P})$communication systems: 1) time and frequency asynchrony of the paired PNC users; and 2) ambiguity of the PNC mapping at the relay node. To address these challenges, in this paper, we apply precoding for power control and phase synchronization of the paired PNC users, while we use modulo$-\sqrt{M}$addition for the PNC mapping ambiguity removal in higher-order M-ary Quadrature Amplitude Modulations (M-QAM). We evaluate the performance of the system in the framework of the Orthogonal Frequency Division Multiplexing (OFDM)-PNC systems with cyclic prefix extension under Rayleigh faded Tapped Delay Line (TDL)-C and Rician faded TDL-D channel models, proposed by the Third Generation Partnership Project$(3\text{GPP})$, as well as the Additive White Gaussian Noise (AWGN) channel model. The results reveal that our proposed technique can achieve a significant Signal-to-Noise Ratio (SNR) improvement of 12$\text{dB}$over its asynchronous OFDM-PNC counterpart (without precoding) for Binary Phase Shift Modulation (BPSK) under a TDL-C faded channel model. Moreover, without channel coding, our proposed PNC technique requires an SNR of around$13\text{dB}$to deliver a two-way 16-QAM communication at a Bit Error Rate of 10−3under a Rayleigh faded TDL-C channel.

Network virtualization allows the coexistence of multiple network slices over a shared physical infrastructure, each delivering a service with own requirements in terms of quality of service, coverage, and time span. Spectrum is an expensive commodity, so it must be managed among these slices in the most efficient way. However, current spectrum sharing techniques are too rigid to address all combinations of service requirements and properly exploit the new air interface flexibility and network deployment options introduced by the Fifth Generation (5G) mobile networks. In this paper, we extend the state of the art first by proposing a novel spectrum sharing scheme that supports an unlimited number of radio networks. Second, we propose a 5G compliant novel service-based network management architecture to integrate Machine Learning (ML) algorithms for Radio Resource Management (RRM), including spectrum sharing. Finally, we propose a new three-stage ML framework that exploits forecasting, clustering and reinforcement learning algorithms to implement the proposed spectrum sharing scheme. Our proposed solution can allow an arbitrary number of 5G network slices to share spectrum more effectively and more dynamically either with each other or with other radio networks.

In the emerging industrial Internet of Things (IIoT) era, machine-to-machine (M2M) communication technology is considered as a key underlying technology for building IIoT environments, where devices (e.g., sensors, actuators, and gateways) are enabled to exchange information with each other in an autonomous way without human intervention. However, most of the existing M2M protocols that can be also used in the IIoT domain provide security mechanisms based on asymmetric cryptography resulting in high computational cost. As a consequence, the resource-constrained IoT devices are not able to support them appropriately and thus, many security issues arise for the IIoT environment. Therefore, lightweight security mechanisms are required for M2M communications in IIoT in order to reach its full potential. As a step toward this direction, in this paper, we propose a lightweight authentication mechanism, based only on hash and XOR operations, for M2M communications in IIoT environment. The proposed mechanism is characterized by low computational cost, communication, and storage overhead, while achieving mutual authentication, session key agreement, device's identity confidentiality, and resistance against the following attacks: replay attack, man-in-the-middle attack, impersonation attack, and modification attack.

Mobile devices are used for communication and for tasks that are sensitive and subject to tampering. Indeed, attacks can be performed on the users' devices without user awareness, this represents additional risk in mission critical scenarios, such as Public Protection and Disaster Relief (PPDR). Intrusion Detection Systems are important for scenarios where information leakage is of crucial importance, since they allow to detect possible attacks to information assets (e.g., installation of malware), or can even compromise the security of PPDR personnel. HyIDS is an Hybrid IDS for Android and supporting the stringent security requirements of PPDR, by comprising agents that continuously monitor mobile device and periodically transmit the data to an analysis framework at the Command Control Center (CCC). The data collection retrieves resource usage metrics for each installed application such as CPU, memory usage, and incoming and outgoing network traffic. At the CCC, the HyIDS employs Machine Learning techniques to identify patterns that are consistent with malware signatures based on the data collected from the applications. The HyIDS's evaluation results demonstrate that the proposed solution has low impact on the mobile device in terms of battery consumption and CPU/memory usage.

An Intrusion Detection System (IDS) aims at protecting a network against attacks intended to exposing and/or vandalizing it. To build and test an IDS, network data are usually acquired containing attacks and normal behavior. The objective of this work is to use machine learning techniques to build IDSs and to investigate their reliability. To build and test the IDSs, KDDCUP99 has been used. The data contain a training set and a testing set with 4,898,430 samples (∼700MB) and 311,032 samples (∼45MB), respectively. However, the cleaned dataset via using SQL commands show that KDDCUP99 is highly redundant. The cleaned/distinct data are nearly one fifth of the original. Subsequently, experimental results have been performed using neural networks based IDSs. Some IDSs give low and median performances when tested using the redundant data and the distinct data, respectively, but other IDSs gave high and median performances using the redundant and the distinct data, respectively. Thus, there is a fluctuation in the performance when the data are redundant, which shows that an IDS built using a redundant dataset has unstable performance. The goal of preparing a balanced dataset is to only use it in testing the realistic performance of the IDS and has no relation to IDS generalization and implementation in real-world scenarios.

Smart grid introduces a wealth of promising applications for upcoming fifth-generation mobile networks (5G), enabling households and utility companies to establish a two-way digital communications dialogue, which can benefit both of them. The utility can monitor real-time consumption of end users and take proper measures (e.g., real-time pricing) to shape their consumption profile or to plan enough supply to meet the foreseen demand. On the other hand, a smart home can receive real-time electricity prices and adjust its consumption to minimize its daily electricity expenditure, while meeting the energy need and the satisfaction level of the dwellers. Smart Home applications for smart phones are also a promising use case, where users can remotely control their appliances, while they are away at work or on their ways home. Although these emerging services can evidently boost the efficiency of the market and the satisfaction of the consumers, they may also introduce new attack surfaces making the grid vulnerable to financial losses or even physical damages. In this paper, we propose an architecture to secure smart grid communications incorporating an intrusion detection system, composed of distributed components collaborating with each other to detect price integrity or load alteration attacks in different segments of an advanced metering infrastructure.

Security is of paramount importance for vehicular communications to verify regular updates from authorized vehicles while detecting misleading reports from malicious or malfunctioning on board units (OBUs). This should however be provided without violating the privacy of the users. Using frequently changing pseudonyms for the vehicles is an effective solution to preserve their privacy. However, in current systems the vehicles need to be always connected to the Certification Authority (CA) to receive the pseudonyms. This might be unfeasible due to the sparse infrastructure deployment, especially in the early stage of the technology, that can lead to availability and scalability problems. In this paper, we solve this issue by proposing efficient protocols for authentication, pseudonym generation, message verification and anonymity revocation that do not require permanent contact with the CA. Even in case of network availability, this feature also reduces the communication overhead for the pseudonym generation and revocation list distribution.

This paper addresses a game-theoretic based distributed appliance scheduling approach for demand-side management in smart grid. The game is played by a utility company and its subscribed residential users. The company monitors electricity demand online, manipulates the price of electricity for different hours to shape users' consumption pattern, and advertises it through Smart Grid infrastructure. Monitoring the electricity price in the market, users independently react by optimally scheduling their shiftable appliances to minimise their electricity expenses. We study day-ahead and quadratic pricing tariffs and discuss the best strategic response of users to each of these strategies that achieves Nash Equilibrium. The game serves as a distributed optimisation tool to minimise load variation in the grid. That is, even if users selfishly minimise their electricity expenses, they will automatically contribute to minimise peak-to-average ratio (PAR) of the aggregate demand, too. Simulation results confirm that adopting quadratic pricing tariff by company and quadratic programming scheduling by users is a Pareto-optimal strategy profile that can achieve a drastic PAR shaving while cutting users electricity expenses by half.

Multi-standard mobile terminals (MTs) allow mobile users to experience ubiquitous connectivity and better quality of service. However, these advances come at a price of higher power consumption for MTs due to holding multiple active interfaces. In this paper, we apply multihop relaying through pervasive short range interfaces (e.g., Bluetooth, WiMedia. etc.) in the presence of an infrastructure network (e.g., LTE, WiFi, etc.) to extend the battery lifetime of multi-standard MTs. To this end, MTs exchange context information (conveying channel state information, battery level, etc.) and, whenever beneficial, form a coalition, pool their resources, and perform their tasks cooperatively to enhance their efficiency. We introduce a novel utility function to assess the profitability of cooperation. To incentivize MTs to cooperate, the achieved common utility of the coalition can be distributed among the cooperative MTs by means of energy credits using a solution concept from coalitional game theory. The simulation results validate the effectiveness of the proposed approach to extend the battery lifetime of MTs to more than double.

Packet forwarding is an essential service in a multihop cellular network (MCN), which relies on the cooperation of all participating mobile terminals (MTs). Although this service can effectively improve the energy efficiency of MTs by dividing a long radio link to several shorter hops, cooperation of MTs cannot be taken for granted as they are normally controlled by rational players who seek to maximize their own benefit by taking advantage of the service while minimizing their contribution. To ensure proper operation of the network, there should be a mechanism to encourage cooperative nodes while punishing free riders. To this end, we propose a credit scheme based on coalitional game model. We provide credit to cooperative nodes proportional to the core solution of the game, which distributes the common utility among the players in a way that everyone is satisfied. Simulation results validate that the proposed technique can successfully detect and isolate selfish nodes.

Multi-standard mobile devices are allowing users to experience higher data rates and ubiquitous connectivity. These advances are achieved on the expense of higher energy consumption due to the multiple active wireless interfaces. In this paper, we use one advantage of the multiple interfaces, namely short range communications. Mobile terminals (MTs) form short range cooperative network to take advantage of the good channel quality of short range links to save energy of MTs. In this cooperative network, the energy of all MTs is treated as a pool of resources, which is available for all MTs in the network. Toward this end, we propose a combinatorial optimization game to identify the problem. We derive the core solution for the relay selection game in cooperative networks. A mathematical framework to compute the achieved energy saving is derived. Simulation results using the solution of the game show that energy saving can be achieved using short range cooperation in wireless networks. The results also show the effectiveness of the proposed node selection game.