Several industries use wireless sensor networks (WSN) for various tasks such as monitoring, data transmission, and data gathering. They find applications in the industrial internet of things (IIoT). WSNs are utilized to track and monitor changes in the environment. Since they include multiple small sensor nodes (SN), they are severely constrained, so resource management geared toward energy efficiency is crucial in this kind of network. Minimizing the power to interpret, transmit, and store data between various sensors poses important challenges. Experts have considered various ways to address these issues that unavoidably affect the network’s performance: reducing energy usage while maintaining system throughput remains the primary research issue. Another important concern relates to network security. Specifically, intrusion detection and avoidance are major concerns. In this work, we introduce the meta-heuristic-based secure and energy-efficient routing (MHSEER) protocol for WSN-IIoT. The protocol learns the forwarding decisions using the number of hops, connection integrity characteristics, and accumulated remaining energy. To make the method more secure, the protocol also employs counter-encryption mode (CEM) to encrypt the data. A meta-heuristics study designed to achieve reliable learning is used in the suggested protocol. The protocol consists of two stages. The first stage uses a heuristics method to improve the option for dependable data routing. Security based on a computationally simple and random CEM is accomplished in the second stage. The proposed MHSEER protocol has been compared to the secure trust routing protocol for low power (Sectrust-RPL), heuristic-based energy-efficient routing (HBEER), secure and energy-aware heuristic-based routing (SEHR), and secure energy-aware meta-heuristic routing (SEAMHR) in terms of packet drop ratio, throughput, network delay, energy usage, and faulty pathways. The proposed protocol increases throughput to 95.81% and decreases the packet drop ratio, packet delay, energy consumption, and faulty pathways to 5.12%, 0.10 ms, 0.0102 mJ, and 6.51%, respectively.

During the recent era, due to the exponential rise in the Internet of things (IoT) enabled devices around the globe, IoT and Machine learning (ML) has come out as usable and efficient approach to supply productive solutions in this environment. Vehicular networks (VN) are considered to be the most important application domain where ML-based techniques are generated to address common attack issues and present insightful discussions. It is the union of smart transport and internet systems which is responsible for the passenger's safety and security as the attack threats have been growing very rapidly in VN. The main objective is to overwhelm the targeted IoT devices with malicious data traffic in VN. To solve the above-mentioned issues, this paper covers numerous types of attacks (backdoor, injection, Distributed Denial of service (DDoS), ransomware, password, scanning, Man in the middle (MITM) and cross-site scripting (XS S)) on VN deploying ML techniques using intrusion detection system (IDS) in the IoT based on the TON-IoT dataset. Due to the fact that it contains a variety of normal and malicious activities for various IoT services as well as heterogeneous data suppliers, TON-IoT provides a number of advantages that are completely missing from state-of-the-art datasets. The performance metrics are evaluated for the attack detection namely accuracy, precision, recall and F1-score using the three ML methods Random Forest (RF), Naive Bayes (NB) and K-Nearest Neighbour (KNN). Finally, the results conclude that amongst the three ML methods, KNN gives the highest accuracy rate i.e. 98.2%, while RF and NB give a 94% and 70% accuracy rate utilizing the ToN-IoT dataset.

Internet of Things (IoT) system is facing a large number of attacks nowadays. Distributed Denial of Ser-vices(DDoS) attack is the most reported attack in the field of security. Nowadays these attacks are increasing very rapidly. However, it has been difficult to predict and detect these attacks easily. In this paper, various existing machine learning (ML) algorithms are being analyzed that are used to predict and detect DDoS attacks using different existing datasets namely; ToN-IoT, CICDDoS2019 and BoT-IoT. The four types of different ML algorithms have been deployed which include K-Nearest Neighbour (KNN), Naive Bayes(NB), Decision Tree (DT) and Random Forest (RF). The major goal of this analysis is to choose the best ML technique. Comparing the existing BoT-IoT dataset to the other datasets utilised in this paper, provides the best measure. The evaluation findings demonstrate that the decision tree and random forest classifiers provide the highest levels of accuracy. The purpose of this paper is to discuss the security concerns associated with DDoS attacks and their mitigation.

The fast growth of 6G-enabled Internet of Things (IoT) networks has transformed communication and made it possible for smart cities, driverless cars, healthcare, and industrial automation to all have seamless connectivity. However, the growing complexity and diversity of 6G-IoT infrastructures present serious cybersecurity issues, leaving these networks open to a range of attacks like malware dissemination, spoofing, and distributed Denial-of-Service (DDoS). Because traditional intrusion detection systems (IDS) cannot adjust to changing attack patterns, they are unable to effectively combat these threats. Using random forest (RF), support vector machine (SVM), logistic regression (LR), and extreme gradient boosting (XGBoost) to improve detection accuracy and robustness, this study proposed a novel Stacked-Hybrid Model for attack detection in 6G-IoT networks. The proposed model maximizes classification performance while reducing computational cost by utilizing feature selection and ensemble learning. Training and evaluation are conducted on the RT-IoT dataset, proving the effectiveness of the method. The experimental results demonstrate that the proposed Stacked-Hybrid Model outperforms individual machine learning (ML) models, achieving 99.90% detection accuracy. The other performance metrics of the proposed model have also been evaluated, including precision, sensitivity, specificity and F1-score rates at 99.56%, 99.56%, 99.93%, and 99.56%, respectively. Significant advancements in identifying intricate and dynamic cyber threats in 6G-IoT contexts are also revealed by a comparison with other models.

Ransomware is a well-known virus variant that has become increasingly prevalent because of the severe and long-lasting damage it causes to its users. Because ransomware causes irreparable damage, quick detection of these threats is essential. Internet of Things(IoT) networks are the main focus of some of the investigations for the security of the system. Over the past ten years, ransomware intrusions have increased, greatly upsetting enterprises. It is essential to create fresh and improved methods for identifying this kind of virus. In order to recognise the constantly changing ransomware attacks in the IoT environment, the research uses several machine learning(ML) approaches. The paper also includes a description of the dataset that was collected from its sources and used for ransomware detection. The study compares the capabilities of various ML techniques for accurately detecting ransomware attacks in IoT systems. The “Cybersecurity: BookMyShow add URL” dataset has been used to apply Logistic Regression(LR), K-Nearest Neighbour(KNN), and eXtreme Gradient Boosting(XGB) as ML techniques. For the purpose of comparing the aforementioned ML approaches, performance metrics including accuracy, recall, precision, and f1-score have been determined. In comparison to LR and KNN, which had accuracy rates of 92.92% and 94.39%, respectively, the system discovered that XGB had the best accuracy, with a score of 96.62%.

This Paper uses the CTU-13 dataset, a large and varied database containing network traffic events, to investigate the use of machine learning(ML) methods for the detection of cyberattacks. The goal is to create reliable intrusion detection models that can precisely locate and categorise instances of both benign and malevolent behaviour in network data. With the use of well-known techniques including Extreme Gradient Boosting (XGBoost), Adaptive Gradient Boosting (AdaBoost), Logistic Regression(LR) and Naive Bayes(NB), the paper attempts to evaluate these models' effectiveness in the environment of cyber threat detection. Accuracy, precision, recall, and F1-score are among the evaluation measures that offer a thorough grasp of the models' efficacy. This study provides insights into the advantages and disadvantages of each algorithm by methodical testing, ad-justing parameters, and cross-validation. It also provides helpful suggestions for choosing the best infrastructure for improving security using machine learning-based threat detection. The results are intended to guide the creation of flexible and potent defence systems against changing cyber threats in actual network environments. According to the data, With an accuracy rate of 96%, the XGBoost model outperforms the other models (NB, LR and AdaBoost) with accuracy rates of 78%, 85%, and 90%, respectively.