Recently, the rapid expansion of the Internet of Things (IoT) has opened up new possibilities and introduced significant security challenges. This evolution enhances everyday life but also increases risks in various domestic and industrial contexts due to growing threats such as cyberattacks and intrusions. To protect both domestic activities and industrial infrastructures, it is imperative to address these challenges. This study enhances security in IoT and IIoT by exploring machine learning-based intrusion detection techniques. The primary goal is to strengthen system protection and ensure the continuity of essential operations. Utilizing the N-BaIoT dataset, designed to simulate realistic IoT attack scenarios, we evaluated the effectiveness of various multiclass classification methods, including PCA dimensionality reduction. After extensive data preprocessing and the application of several classifiers such as KNN, Random Forest, Naive Bayes, Decision Tree, Extra Trees, and XGBoost, we built an effective IoT IDS. The Extra Trees algorithm, in combination with PCA, showed the best performance, achieving an impressive 99.94% accuracy. This underscores the effectiveness of machine learning in detecting and mitigating IoT and IIoT cyber threats and highlights the importance of selecting appropriate methods for optimal results in complex security environments.

In critical IoT environments, such as smart homes and industrial systems, effective Intrusion Detection Systems (IDS) are essential for ensuring security. However, developing robust IDS solutions remains a significant challenge. Traditional machine learning-based IDS models typically require large datasets, but data sharing is often limited due to privacy and security concerns. Federated Learning (FL) presents a promising alternative by enabling collaborative model training without sharing raw data. Despite its advantages, FL still faces key challenges, such as data heterogeneity (non-IID data) and high energy and computation costs, particularly for resource-constrained IoT devices. To address these issues, this paper proposes OptiFLIDS, a novel approach that applies pruning techniques during local training to reduce model complexity and energy consumption. It also incorporates a customized aggregation method to better handle pruned models that differ due to non-IID data distributions. Experiments conducted on three recent IoT IDS datasets, TON_IoT, X-IIoTID, and IDS-IoT2024, demonstrate that OptiFLIDS maintains strong detection performance while improving energy efficiency, making it well-suited for deployment in real-world IoT environments.

Intrusion Detection Systems (IDSs) have played a significant role in detecting and preventing cyber-attacks within traditional computing systems. It is not surprising that the same technology is being applied to secure Internet of Things (IoT) networks from cyber threats. The limited computational resources available on IoT devices make it challenging to deploy conventional computing-based IDSs. The IDSs designed for IoT environments must also demonstrate high classification performance, utilize low-complexity models, and be of a small size. Despite significant progress in IoT-based intrusion detection, developing models that both achieve high classification performance and maintain reduced complexity remains challenging. In this study, we propose a hybrid CNN architecture composed of a lightweight CNN and bidirectional LSTM (BiLSTM) to enhance the performance of IDS on the UNSW-NB15 dataset. The proposed model is specifically designed to run onboard resource-constrained IoT devices and meet their computation capability requirements. Despite the complexity of designing a model that fits the requirements of IoT devices and achieves higher accuracy, our proposed model outperforms the existing research efforts in the literature by achieving an accuracy of 97.28% for binary classification and 96.91% for multiclassification.

With the rapid growth in the Internet of Things (IoT), current cybersecurity threats are growing to levels at which traditional intrusion detection systems (IDSs) cannot suffice. The purpose of this paper, therefore, is to evaluate the efficacy of ML techniques in enhancing IDS to adapt to emerging sophisticated and dynamic modern cyber threats. We adopted four supervised ML models: Decision Tree, Random Forest, Naive Bayes, and Gradient Boost, all of which describe the CSE-CIC-IDS2018 dataset representing different network attack situations. This reflects today's cyber threats. We have analyzed both binary and multiclass classification tasks to understand what kind of cyber threat and how many any model was best suitable for. This implies that ML-supported IDS can effectively improve the detection of not only generic but also specific cyber threats, therefore enhancing security. It is in line with this that the current paper outlines the strengths and weaknesses of the discussed models to produce an insight and judgment of their practical implementation in real-world scalability. The outcomes came into view that ML-based IDS gives resilient, adaptable, and proactive solutions to cybersecurity over IoT networks.

Intrusion Detection Systems (IDSs) have played a significant role in the detection and prevention of cyber-attacks in traditional computing systems. It is not surprising that this technology is now being applied to secure Internet of Things (IoT) networks against cyber threats. However, the limited computational resources available on IoT devices pose a challenge for deploying conventional computing-based IDSs. IDSs designed for IoT environments must demonstrate high classification performance, and utilize low-complexity models. Developing intrusion detection models in the field of IoT has seen significant advancements. However, achieving a balance between high classification performance and reduced complexity remains a challenging endeavor. In this research, we present an effective IDS model that addresses this issue by combining a lightweight Convolutional Neural Network (CNN) with bidirectional Long Short-Term Memory (BiLSTM). Additionally, we employ feature selection techniques to minimize the number of features inputted into the model, thereby reducing its complexity. This approach renders the proposed model highly suitable for resource-constrained IoT devices, ensuring it meets their computation capability requirements. Creating a model that meets the demands of IoT devices and attains enhanced precision is a challenging task. However, our suggested model outperforms previous works in the literature by attaining a remarkable accuracy rate of 97.90% within a prediction time of 1.1 seconds for binary classification. Furthermore, it achieves an accuracy rate of 97.09% within a prediction time of 2.10 seconds for multiclassification.

Wireless communication systems rely on channel estimation and equalization to ensure reliable and efficient data transmission. However, with the increasing demand for high connectivity in massive IoT networks, these processes are facing significant challenges. The complexity and intensive computation required for channel estimation and equalization results in high communication latency and power consumption, which can ultimately prevent the transceiver from restoring the originally transmitted signal. In this paper, we propose a novel approach to simplify wireless communication systems by using a conditional generative adversarial network (cGAN) model to replace both channel estimation and equalization blocks. We formulate the data recovery task as a translation from received data to the corresponding transmitted signal and introduce the concept of Rx-to-Tx translation based on a cGAN, which was initially developed for image-to-image translation. Our preliminary results demonstrate the feasibility and effectiveness of this approach, particularly for digital modulations. By carefully tuning the model's hyperparameters, we achieve the theoretical symbol error rate (SER) of QAMs in a Rayleigh propagation channel. Our proposed approach has the potential to significantly reduce the computational complexity and overhead typically associated with traditional channel estimation and equalization blocks. This can lead to more efficient and cost-effective wireless communication systems.

In the last few years, there has been a massive increase in Internet of Things (IoT) devices and the data generated from these appliances. Devices involved in IoT networks can be challenging because of their resource-constrained nature, and security integration's on these devices are frequently disregarded. This results in attackers targeting more IoT devices. Thus, as the number of possible attacks on a network increases, it becomes more difficult for traditional intrusion detection systems (IDS) to deal with these attacks effectively. This paper presents a hybrid deep learning-based approach, a one- dimensional convolutional neural network, and long short-term memory (1D CNN-LSTM) algorithm, for anomaly detection that harnesses the power of the IoT, providing qualities to efficiently examine all traffic across the IoT. The comprehensive study was conducted utilizing the Bot-IoT dataset extracted from real network traffic, consisting of benign and malicious variants. Then, the anomaly detection including binary and multi-decision categories has been performed. The experimental results highlighted the superiority of the proposed model with an accuracy of 99.20% and lower false alarm with 0.80% compared to single CNN-based IDS.

A resource-constrained unmanned aerial vehicle (UAV) can be used as a flying LoRa gateway (GW) to move inside the target area for efficient data collection and LoRa resource management. In this work, we propose deep reinforcement learning (DRL) to optimize the energy efficiency (EE) in wireless LoRa networks composed of LoRa end devices (EDs) and a flying GW to extend the network lifetime. The trained DRL agent can efficiently allocate the spreading factors (SFs) and transmission powers (TPs) to EDs while considering the air-to-ground wireless link and the availability of SFs. In addition, we allow the flying GW to adjust its optimal policy onboard and perform online resource allocation. This is accomplished through retraining the DRL agent using reduced action space. Simulation results demonstrate that our proposed DRL-based online resource allocation scheme can achieve higher EE in LoRa networks over three benchmark schemes.

Most of the developed localization solutions rely on RSSI fingerprinting. However, in the LoRa networks, due to the spreading factor (SF) in the network setting, traditional fingerprinting may lack representativeness of the radio map, leading to inaccurate position estimates. As such, in this work, we propose a novel LoRa RSSI fingerprinting approach that takes into account the SF. The performance evaluation shows the prominence of our proposed approach since we achieved an improvement in localization accuracy by up to 6.67% compared to the state-of-the-art methods. The evaluation has been done using a fully connected deep neural network (DNN) set as the baseline. To further improve the localization accuracy, we propose a deep reinforcement learning model that captures the ever-growing complexity of LoRa networks and copes with their scalability. The obtained results show an improvement of 48.10% in the localization accuracy compared to the baseline DNN model.

During the last decade, the integration of artificial intelligence (AI) and the use of intrusion detection systems (IDSs) in the Internet of Things(IoT) networks have brought a new dimension to technological progress. Deep learning (DL) and machine learning (ML)-based IDS are vulnerable to adversarial perturbations. However, anomaly detection methods suffer from unbalanced and missing sample data, thus causing IDS training to be complicated. In this paper, we propose using conditional generative adversarial networks (cGANs) to enhance the training process by handling the unbalanced data and coping with the lack of specifics class samples, which may succeed in evading our Convolutional Neural Network-Long Short-Term Memory (CNNLSTM) based-IDS model. We evaluated our proposed IDS model before and after applying the adversarial training using the Bot-IoT dataset. Promising results showed that the accuracy of detecting Theft attacks could be increased by 40%. To the best of our knowledge, we are the first to suggest the combination of cGAN and CNNLSTM based-IDS system to enhance its performance.

Selfish mining is an attack on blockchain networks, where a minority mining pool deviates from the original mining protocol and keeps some blocks private. The goal of the attacking pool is to waste the computational power of the other miners and increase their revenue. In this paper, we use a new approach to analyze the profitability of such attacks. Using game theory, we model the interactions between pools to derive the utility of mining strategies. We simulate the game for a Bitcoin blockchain and analyze the profitability of an attack, in terms of the monetary award instead of the relative revenue. We express the utility to include the cost of a strategy and revisit existing selfish mining strategies to discuss possible outcomes of the game. Depending on the game parameterization, we highlight scenarios where the system could be compromised. To the best of our knowledge, this is the first work that models the selfish mining attack as a stochastic game.

Unmanned aerial vehicles (UAVs) have attracted great interest in the last few years owing to their ability to cover large areas and access difficult and hazardous target zones, which is not the case of traditional systems relying on direct observations obtained from fixed cameras and sensors. Furthermore, thanks to the advancements in computer vision and machine learning, UAVs are being adopted for a broad range of solutions and applications. However, deep neural networks (DNNs) are progressing toward deeper and complex models that prevent them from being executed onboard. In this article, we propose a DNN distribution methodology within UAVs to enable data classification in resource-constrained devices and avoid extra delays introduced by the server-based solutions due to data communication over air-to-ground links. The proposed method is formulated as an optimization problem that aims to minimize the latency between data collection and decision-making while considering the mobility model and the resource constraints of the UAVs as part of the air-to-air communication. We also introduce the mobility prediction to adapt our system to the dynamics of UAVs and the network variation. The simulation conducted to evaluate the performance and benchmark the proposed methods, namely, optimal UAV-based layer distribution (OULD) and OULD with mobility prediction (OULD-MP), was run in an HPC cluster. The obtained results show that our optimization solution outperforms the existing and heuristic-based approaches.

With considerable demand for higher throughput, greater capacity, and lower latency for consumers, the Internet of Things (IoT) network is anticipated to meet the desired security and privacy requirements. This study provides high transaction throughput on critical IoT applications, particularly Bitcoin security against double-spending attacks. To this end, we investigated the signaling game approach to model the interaction between two miners while considering players behavior (malicious or honest miners) and the incoming transaction throughput. To the best of our knowledge, this is the first work that exploits the signaling game to cover the incoming transactions randomness waiting for validation, which influences the honest miners behavior. With extensive simulations, we show that our proposed signaling game reduces the impact of double-spending attacks on IoT transactions. The results also illustrate the benefit of using the signaling game to model the interaction between two miners while handling the incomplete information of the incoming transactions and the type of miners.

Underwater Sensor Network (UWSN) suffers from the limited batteries life of sensor nodes. Thus, some nodes will disappear from the network topology during the communication process which leads to isolated nodes and important buffered packets will be discarded. Traditional greedy forwarding protocol used in UWS N s are based on the selection of the nearest next-hop forwarder from the destination, that's the nearest one from the sea surface relaying the source and destination. By this, some nodes are selected by multiple source nodes, so their energy risk to be drained. In order to overcome this problem, we enhance this protocol by distributing the forwarding task between multiple next-hop forwarders. Also, this protocol is based on depth adjustment to solve the problem of isolated nodes. The source packet is splitted and each sub-packet is transmitted to a single upper neighbor node. Otherwise, multiple data channels are used to avoid collision between source nodes selecting the same next-hop node. Numerical results show significant improvement in greedy forwarding protocol performance.

Recently an extensive deployment of small cell Networks (SCNs) called two tier-heterogeneous networks have been proposed, especially for dense urban zone. Therefore, Macro Cell Networks containing a hyper density of SCNs distributed randomly within it will emerge the field. The benefits of two tier Heterogeneous Networks (HetNets) are realized through the traffic flooding via small cells. To enhance the performance of the two-tier cellular networks and offer the best quality service for Macro user, the cell association problem is modeled using the partially observable Markov decision process (POMDP). POMDP works based on a set of beliefs states that the macro-user get considering the impact of the base station's (BS) best channel. The macro-user has to make decisions based on its channel state partially observable and the shared information in order to associate with BS, the association decisions are determined based on the calculated SINR. The numerical results of the proposed solution are given with comments and they show the acceptable system performance.

Recently Underwater wireless Sensor Networks (UWSNs) have been suggested as a powerful technology for many civilian and military applications, such that tactical surveillance. Geographic routing that uses the position information of nodes to route the packet toward a destination is preferable for UWSNs. In this paper, we propose a New Greedy Forwarding (NGF) strategy using splitting mechanism based on Chinese remainder theorem(CRT) for UWSNs. In the approach, source node reduced the number of bits transmitted using the proposed splitting mechanism based on CRT if there are more than two nodes participate in the forwarding of one packet. This strategy distribute the forwarding task between more nodes instead of selecting one node as next-hop, that reduce the energy consumption per node and maintain node communication for a long time. Thus resulting the increase of network life time and decrease the number of isolated/void nodes. We use topology control through depth adjustment to cope with the problem of isolated and void nodes appeared in geographic routing protocols. Simulation results shows that with the anycast greedy forwarding strategy the network life time is about 500 rounds whereas it is about 1000 rounds using the new greedy forwarding strategy, which means the new strategy increase the network life time and increase the network performance in energy saving.

Recently Underwater Sensor Networks (UWSNs) have been suggested as a powerful technology for many civilian and military applications, such as tactical surveillance. The most important issue in these networks is the communication, mainly due to the presence of fading, multi-path and refractive properties of the sound channel, this necessitate the development of precise underwater channel model for each application and provide an efficient routing protocol that consider the energy constraint of underwater nodes and resolve the problem of disconnected nodes. In this work, we study the impact of two topology control methods, that are used to resolve the problem of void/isolated nodes appeared in geographic routing protocols, in network performance. Simulation results of topology control through Depth adjustment DA and Transmission Power Controls TPC showed a significant reduction of the number of void/isolated nodes.