دانلود کتاب کاربرد هوش مصنوعی در عصر جدید شبکه‌های ارتباطی

The applications of machine learning in wireless and mobile communication net-works have been receiving increasing attention, especially in the new era of big data and the Internet of Things (IoT), where data mining and data analysis technologies are effective approaches to solving wireless system issues. Artificial intelligence is one of the leading technologies in 5G, beyond 5G, and future 6G networks. Intelligence is playing a crucial role in unlocking the full potential of the 5G networks and the future 6G mobile wireless networks by leveraging universal infrastructure, open network architectures, software-defined networking, network function virtualization, multi-access edge computing, vehicular networks, etc. The implementation of blockchain and mobile edge computing have become a significant part of the new wireless and mobile communication networks, helping to perform computations as close to IoT devices as possible. This Special Issue offers valuable contributions to wireless and mobile communication technologies, mobile edge computing, and blockchain, using modern artificial intelligence and machine learning techniques. We received over 24 submissions. After rigorous manuscript screening and a peer-review process, ten articles were accepted for this Special Issue. In the following paragraphs, we provide summaries of these contributions. Contribution 1 carries out traffic classification in a software-defined network (SDN)/cloud environment using supervised learning with four different algorithms—Naive Bayes, Support Vector Machines (SVMs), Random Forest and J48 tree (C4.5). The authors discuss the methodology used based on the Weka tool and explain how the testbed was deployed using two sets of features. The goal of Contribution 2 is to create Quantized Neural Network (QNN) models to efficiently classify rice varieties while reducing resource usage on edge devices. The authors state that QNN is an effective method for alleviating high computational costs and power requirements in response to many deep learning (DL) algorithms. They also create/develop eight different QNNs using MLP and LeNet-5-based deep learning models with varying quantization levels to be trained by the dataset. In Contribution 3, task-related adaptation on the machine side is proposed to account for a person’s momentous cognitive and emotional state. The authors validate the practical significance of the proposed approach in both person-specific and person-independent setups. The analysis of results in the person-specific setup shows that the individual optimal performance curves for that person, according to the Yerkes–Dodson law, are dis-placed. In Contribution 4, the authors presents a comprehensive research study on integrating software-defined networking (SDN) with machine learning (ML) to improve network performance and Quality of Service (QoS). The study primarily investigates ML classification methods, highlighting their significance in the context of traffic classification (TC). Traditional methods are discussed to clarify the observed outperformance of ML throughout the investigation, underscoring the superiority of ML algorithms in SDN TC. Contribution 5 proposes a method to generate adversarial samples of frequencyhopping communication signals using adaptive whitening and feature gradient smoothing. The proposed method targets the DNN cognitive link of the jammer, aiming to reduce modulation recognition accuracy and counteract smart interference. In Contribution 6, the authors state that the detection of MC-SDN security attacks is crucial for improving network performance. For this reason, they designed blockchainbased controller security (BCS), which exploits the advantages of immutable and distributed ledger technology among multiple controllers and securely manages controller communications against various attacks. The experimental results of the proposed BCS approach demonstrate its performance under various scenarios, such as attack detection, number of attackers, number of controllers, and number of compromised controllers, by applying different performance metrics. The objective of Contribution 7 is to develop a model to improve the stability of a low-cost receiver. To achieve this, the authors propose a machine learning-based linear regression algorithm to predict the differences in the low-cost GPS receiver compared to the precision timing source. Contribution 8 aims to create an autonomous indoor navigation system for medical robots using sensors such as Marvelmind, LiDAR, IMU, and an odometer, along with the Time Elastic Band (TEB) local planning algorithm to detect dynamic obstacles. The algorithm’s performance is evaluated using metrics such as path length, duration, speed smoothness, path smoothness, Mean Squared Error (MSE), and positional error. In the test arena, TEB demonstrates superior efficiency with a path length of 155.55 m, 9.83 m shorter than the DynamicWindow Approach (DWA), which covered 165.38 m, and a lower yaw error of 0.012 radians. Contribution 9 presents an Application-Specific Integrated Circuit (ASIC) implementation and Field-Programmable Gate Array (FPGA) verification of a Convolutional Neural Network (CNN)-based Intrusion Detection System (IDS) designed to enhance the security of an in-vehicle Controller Area Network (CAN) bus and detect malicious messages. In this article, the authors also conduct functional verification by configuring a simplified CAN bus environment using the Xilinx Nexys Video FPGA and PEAK-System PCAN-USB, which is validated in real time against DoS, spoofing, and fuzzy attack scenarios. The YOLO-based object detection model is used in Contribution 10 to identify pedestrians and extract key data such as bounding box coordinates and confidence levels. These data are encoded afterward into decentralized environmental notification messages (DENMs) using ASN.1 schemas to ensure compliance with V2X standards, allowing for real-time communication between vehicles and infrastructure. The authors identify that the integration of pedestrian detection with V2X communication resulted in a reliable system wherein the roadside unit (RSU) broadcasts DENM alerts to vehicles. Upon receiving the messages, the vehicles initiate appropriate responses, such as slowing down or lane changing, with the testing demonstrating reliable message transmission and high pedestrian detection accuracy in simulated–controlled environments.