As 5G and other technologies are widely used in the Internet of Vehicles, intrusion detection plays an increasingly important role as a vital detection tool for information security. However, due to the rapid changes in the structure of the Internet of Vehicles, the large data flow, and the complex and diverse forms of intrusion, traditional detection methods cannot ensure their accuracy and real-time requirements and cannot be directly applied to the Internet of Vehicles. A new AA distributed combined deep learning intrusion detection method for the Internet of Vehicles based on the Apache Spark framework is proposed in response to these problems. The cluster combines deep-learning convolutional neural network (CNN) and extended short-term memory (LSTM) network to extract features and data for detection of car network intrusion from large-scale car network data traffic and discovery of abnormal behavior. The experimental results show that compared with other existing models, the algorithm of this model can reach 20 in the fastest time, and the accuracy rate is up to 99.7%, with a good detection effect.

The Internet of Medical Things (IoMT) has emerged as one of the most important key applications of IoT. IoMT makes the diagnosis and care more convenient and reliable with proven results. The paper presents the technology, open issues, and challenges of IoMT-based systems. It explores the various types of sensors and smart equipment based on IoMT and used for diagnosis and patient care. A comprehensive survey of early detection and postdetection care of the neural disorder dementia is conducted. The paper also presents a postdiagnosis dementia care model named “Demencare.” This model incorporates eight sensors capable of tracking the daily routine of dementia patient. The patients can be monitored locally by an edge computing device kept at their premises. The medical experts may also monitor the patients’ status for any deviation from normal behavior. IoMT enables better postdiagnosis care for neural disorders, like dementia and Alzheimer’s. The patient’s behavior and vital parameters are always available despite the remote location of the patients. The data of the patients may be classified, and new insights may be obtained to tackle patients in a better manner.

With the rapid increase and complexity of IPv6 network traffic, the traditional intrusion detection system Snort detects DoS attacks based on specific rules, which reduces the detection performance of IDS. To solve the DoS intrusion detection problem in the IPv6 network environment, the lightweight KNN optimization algorithm in machine learning is adopted. First, the double dimensionality reduction of features is achieved through the information gain rate, and discrete features with more subfeatures are selected and aggregated to further dimensionality reduction and feature dimension of the actual operation. Secondly, the information gain rate is used as the weight to optimize the sample Euclidean distance measurement. Based on the proposed measure of the reverse distance influence, the classification decision algorithm of the KNN algorithm is optimized to make the detection technology better. The effect is further improved. The experimental results show that the traditional TAD‐KNN algorithm based on average distance and the GR‐KNN algorithm that only optimizes the distance definition, the GR‐AD‐KNN algorithm can not only improve the overall detection performance in the detection of IPv6 network traffic characteristics but also for small groups of samples. As a result, classification has better detection results.

The Internet of Vehicles (IOV) allows vehicles to communicate with each other in the Internet of Things (IOT). As vehicle nodes are considered to be always in motion, their topology frequently changes. The dynamic topology changes that result in these changes have caused IOV to face major issues, including scalability, shortest-path routing, and dynamic topology changes. Clustering can be used to solve such problems. Clustering is based on an optimization approach based on transmission range, node density, speed, and direction. This paper presents a method for calculating and evaluating an optimal cluster head (CH) using ant colony and Firefly optimization algorithms. Massively interconnected networks with heterogeneous data generated at the edge of networks require distributed machine-learning techniques that can take advantage of this data. A three-layer federated learning model is proposed in this study to take advantage of the distributed end-edge-cloud architecture typical of a 5G/6G environment to increase learning efficiency and accuracy while protecting data privacy and reducing communications overhead. Our experimental and evaluation results demonstrate our proposed method’s outstanding performance in improving convergence speed and learning accuracy for 5G/6G-supported IoV applications. The proposed TFL-IHOA model enhanced the number of clusters in the grid by 3-5%, reduced the computation time by 1.5-2%, and had 12-20% less packet loss than the existing algorithms.