The development of communication technology will promote the application of Internet of Things, and Beyond 5G will become a new technology promoter. At the same time, Beyond 5G will become one of the important supports for the development of edge computing technology. This paper proposes a communication task allocation algorithm based on deep reinforcement learning for vehicle-to-pedestrian communication scenarios in edge computing. Through trial and error learning of agent, the optimal spectrum and power can be determined for transmission without global information, so as to balance the communication between vehicle-to-pedestrian and vehicle-to-infrastructure. The results show that the agent can effectively improve vehicle-to-infrastructure communication rate as well as meeting the delay constraints on the vehicle-to-pedestrian link.

Students' demand for online learning has exploded during the post-COVID-19 pandemic era. However, due to their poor learning experience, students' dropout rate and learning performance of online learning are not always satisfactory. The technical advantages of Beyond Fifth Generation (B5G) can guarantee a good multimedia Quality of Experience (QoE). As a special case of multimedia services, online learning takes into account both the usability of the service and the cognitive development of the users. Factors that affect the Quality of Online Learning Experience (OL-QoE) become more complicated. To get over this dilemma, we propose a systematic scheme by integrating big data, Machine Learning (ML) technologies, and educational psychology theory. Specifically, we first formulate a general definition of OL-QoE by data analysis and experimental verification. This formula considers both the subjective and objective factors (i.e., video watching ratio and test scores) that most affect OL-QoE. Then, we induce an extended layer to the classic Broad Learning System (BLS) to construct an Extended Broad Learning System (EBLS) for the students' OL-QoE prediction. Since the extended layer can increase the width of the BLS model and reduce the redundant nodes of BLS, the proposed EBLS can achieve a trade-off between the prediction accuracy and computation complexity. Finally, we provide a series of early intervention suggestions for different types of students according to their predicted OL-QoE values. Through timely interventions, their OL-QoE and learning performance can be improved. Experimental results verify the effectiveness of the proposed scheme.

Reliable communication and intensive computing power cannot be provided effectively by temporary hot spots in disaster areas and complex terrain ground infrastructure. Mitigating this has greatly developed the application and integration of UAV and Mobile Edge Computing (MEC) to the Internet of Things (IoT). However, problems such as multi-user and huge data flow in large areas, which contradict the reality that a single UAV is constrained by limited computing power, still exist. Due to allowing UAV collaboration to accomplish complex tasks, coop- erative task offloading between multiple UAVs must meet the interdependence of tasks and realize parallel processing, which reduces the computing power consumption and endurance pressure of terminals. Considering the computing requirements of the user terminal, delay constraint of a computing task, energy constraint, and safe distance of UAV, we constructed a UAV-Assisted cooperative offloading energy efficiency system for mobile edge computing to minimize user terminal energy consumption. However, the resulting optimization problem is originally nonconvex and thus, difficult to solve optimally. To tackle this problem, we developed an energy ef- ficiency optimization algorithm using Block Coordinate Descent (BCD) that decomposes the problem into three convex subproblems. Furthermore, we jointly optimized the number of local computing tasks, number of computing offloaded tasks, trajectories of UAV, and offloading matching relationship between multi-UAVs and multiuser terminals. Simulation results show that the proposed approach is suitable for different channel con- ditions and significantly saves the user terminal energy consumption compared with other benchmark schemes.

Unbalanced traffic distribution in cellular networks results in congestion and degrades spectrum efficiency. To tackle this problem, we propose an Unmanned Aerial Vehicle (UAV)-assisted wireless network in which the UAV acts as an aerial relay to divert some traffic from the overloaded cell to its adjacent underloaded cell. To fully exploit its potential, we jointly optimize the UAV position, user association, spectrum allocation, and power allocation to maximize the sum-log-rate of all users in two adjacent cells. To tackle the complicated joint opti- mization problem, we first design a genetic-based algorithm to optimize the UAV position. Then, we simplify the problem by theoretical analysis and devise a low-complexity algorithm according to the branch-and-bound method, so as to obtain the optimal user association and spectrum allocation schemes. We further propose an iterative power allocation algorithm based on the sequential convex approximation theory. The simulation results indicate that the proposed UAV-assisted wireless network is superior to the terrestrial network in both utility and throughput, and the proposed algorithms can substantially improve the network performance in comparison with the other schemes.

Beyond-5G (B5G) aims to meet the growing demands of mobile traffic and expand the communication space. Considering that intelligent applications to B5G wireless communications will involve security issues regarding user data and operational data, this paper analyzes the maximum capacity of the multi-watermarking method for multimedia signal hiding as a means of alleviating the information security problem of B5G. The multi-watermarking process employs spread transform dither modulation. During the watermarking procedure, Gram-Schmidt orthogonalization is used to obtain the multiple spreading vectors. Consequently, multiple wa- termarks can be simultaneously embedded into the same position of a multimedia signal. Moreover, the multiple watermarks can be extracted without affecting one another during the extraction process. We analyze the effect of the size of the spreading vector on the unit maximum capacity, and consequently derive the theoretical rela- tionship between the size of the spreading vector and the unit maximum capacity. A number of experiments are conducted to determine the optimal parameter values for maximum robustness on the premise of high capacity and good imperceptibility.

In this paper, we jointly design the power control and position dispatch for Multi-Unmanned Aerial Vehicle (UAV)-enabled communication in Device-to-Device (D2D) networks. Our objective is to maximize the total transmission rate of Downlink Users (DUs). Meanwhile, the Quality of Service (QoS) of all D2D users must be satisfied. We comprehensively considered the interference among D2D communications and downlink trans- missions. The original problem is strongly non-convex, which requires high computational complexity for traditional optimization methods. And to make matters worse, the results are not necessarily globally optimal. In this paper, we propose a novel Graph Neural Networks (GNN) based approach that can map the considered system into a specific graph structure and achieve the optimal solution in a low complexity manner. Particularly, we first construct a GNN-based model for the proposed network, in which the transmission links and interference links are formulated as vertexes and edges, respectively. Then, by taking the channel state information and the coordinates of ground users as the inputs, as well as the location of UAVs and the transmission power of all transmitters as outputs, we obtain the mapping from inputs to outputs through training the parameters of GNN. Simulation results verified that the way to maximize the total transmission rate of DUs can be extracted effectively via the training on samples. Moreover, it also shows that the performance of proposed GNN-based method is better than that of traditional means.

Unmanned Aerial Vehicles (UAVs) as aerial base stations to provide communication services for ground users is a flexible and cost-effective paradigm in B5G. Besides, dynamic resource allocation and multi-connectivity can be adopted to further harness the potentials of UAVs in improving communication capacity, in such situations such that the interference among users becomes a pivotal disincentive requiring effective solutions. To this end, we investigate the Joint UAV-User Association, Channel Allocation, and transmission Power Control (J-UACAPC) problem in a multi-connectivity-enabled UAV network with constrained backhaul links, where each UAV can determine the reusable channels and transmission power to serve the selected ground users. The goal was to mitigate co-channel interference while maximizing long-term system utility. The problem was modeled as a cooperative stochastic game with hybrid discrete-continuous action space. A Multi-Agent Hybrid Deep Rein- forcement Learning (MAHDRL) algorithm was proposed to address this problem. Extensive simulation results demonstrated the effectiveness of the proposed algorithm and showed that it has a higher system utility than the baseline methods.

Haptic communications is recognized as a promising enabler of extensive services by enabling real-time haptic control and feedback in remote environments, e.g., teleoperation and autonomous driving. Considering the strict transmission requirements on reliability and latency, Device-to-Device (D2D) communications is introduced to assist haptic communications. In particular, the teleoperators with poor channel quality are assisted by auxiliaries, and each auxiliary and its corresponding teleoperator constitute a D2D pair. However, the haptic interaction and the scarcity of radio resources pose severe challenges to the resource allocation, especially facing the sporadic packet arrivals. First, the contention-based access scheme is applied to achieve low-latency transmission, where the resource scheduling latency is omitted and users can directly access available resources. In this context, we derive the reliability index of D2D pairs under the contention-based access scheme, i.e., closed-loop packet error probability. Then, the reliability performance is guaranteed by bidirectional power control, which aims to minimize the sum packet error probability of all D2D pairs. Potential game theory is introduced to solve the problem with low complexity. Accordingly, a distributed power control algorithm based on synchronous log-linear learning is proposed to converge to the optimal Nash Equilibrium. Experimental results demonstrate the superiority of the proposed learning algorithm.

Along with the development of 5G network and Internet of Things technologies, there has been an explosion in personalized healthcare systems. When the 5G and Artificial Intelligence (AI) is introduced into diabetes management architecture, it can increase the efficiency of existing systems and complications of diabetes can be handled more effectively by taking advantage of 5G. In this article, we propose a 5G-based Artificial Intelligence Diabetes Management architecture (AIDM), which can help physicians and patients to manage both acute complications and chronic complications. The AIDM contains five layers: the sensing layer, the transmission layer, the storage layer, the computing layer, and the application layer. We build a test bed for the transmission and application layers. Specifically, we apply a delay-aware RA optimization based on a double-queue model to improve access efficiency in smart hospital wards in the transmission layer. In application layer, we build a prediction model using a deep forest algorithm. Results on real-world data show that our AIDM can enhance the efficiency of diabetes management and improve the screening rate of diabetes as well.

The great potentials of massive Multiple-Input Multiple-Output (MIMO) in Frequency Division Duplex (FDD) mode can be fully exploited when the downlink Channel State Information (CSI) is available at base stations. However, the accurate CSI is difficult to obtain due to the large amount of feedback overhead caused by massive antennas. In this paper, we propose a deep learning based joint channel estimation and feedback framework, which comprehensively realizes the estimation, compression, and reconstruction of downlink channels in FDD massive MIMO systems. Two networks are constructed to perform estimation and feedback explicitly and implicitly. The explicit network adopts a multi-Signal-to-Noise-Ratios (SNRs) technique to obtain a single trained channel estimation subnet that works well with different SNRs and employs a deep residual network to reconstruct the channels, while the implicit network directly compresses pilots and sends them back to reduce network parameters. Quantization module is also designed to generate data-bearing bitstreams. Simulation results show that the two proposed networks exhibit excellent performance of reconstruction and are robust to different environments and quantization errors.

The research for the Intelligent Reflecting Surface (IRS) which has the advantages of cost and energy efficiency has been studied. Channel capacity can be effectively increased by appropriately setting the phase value of IRS elements according to the channel conditions. However, the problem of obtaining an appropriate phase value of IRS is difficult to solve due to the non-convex problem. This paper proposes an iterative algorithm for the alternating optimal solution in the Single User Multiple-Input-Multiple-Output (SU-MIMO) systems. The proposed iterative algorithm finds an alternating optimal solution that is the phase value of IRS one by one. The results show that the proposed method has better performance than that of the randomized IRS systems. The number of iterations for maximizing the performance of the proposed algorithm depends on the channel state between the IRS and the receiver.

Millimeter-wave transmission combined with Orbital Angular Momentum (OAM) has the advantage of reducing the loss of beam power and increasing the system capacity. However, to fulfill this advantage, the antennas at the transmitter and receiver must be parallel and coaxial; otherwise, the accuracy of mode detection at the receiver can be seriously influenced. In this paper, we design an OAM millimeter-wave communication system for overcoming the above limitation. Specifically, the first contribution is that the power distribution between different OAM modes and the capacity of the system with different mode sets are analytically derived for performance analysis. The second contribution lies in that a novel mode selection scheme is proposed to reduce the total interference between different modes. Numerical results show that system performance is less affected by the offset when the mode set with smaller modes or larger intervals is selected.

In Beyond the Fifth Generation (B5G) heterogeneous edge networks, numerous users are multiplexed on a channel or served on the same frequency resource block, in which case the transmitter applies coding and the receiver uses interference cancellation. Unfortunately, uncoordinated radio resource allocation can reduce system throughput and lead to user inequity, for this reason, in this paper, channel allocation and power allocation problems are formulated to maximize the system sum rate and minimum user achievable rate. Since the construction model is non-convex and the response variables are high-dimensional, a distributed Deep Reinforcement Learning (DRL) framework called distributed Proximal Policy Optimization (PPO) is proposed to allocate or assign resources. Specifically, several simulated agents are trained in a heterogeneous environment to find robust behaviors that perform well in channel assignment and power allocation. Moreover, agents in the collection stage slow down, which hinders the learning of other agents. Therefore, a preemption strategy is further proposed in this paper to optimize the distributed PPO, form DP-PPO and successfully mitigate the straggler problem. The experimental results show that our mechanism named DP-PPO improves the performance over other DRL methods.

The performance of massive MIMO systems relies heavily on the availability of Channel State Information at the Transmitter (CSIT). A large amount of work has been devoted to reducing the CSIT acquisition overhead at the pilot training and/or CSI feedback stage. In fact, the downlink communication generally includes three stages, i.e., pilot training, CSI feedback, and data transmission. These three stages are mutually related and jointly determine the overall system performance. Unfortunately, there exist few studies on the reduction of CSIT acquisition overhead from the global point of view. In this paper, we integrate the Minimum Mean Square Error (MMSE) channel estimation, Random Vector Quantization (RVQ) based limited feedback and Maximal Ratio Combining (MRC) precoding into a unified framework for investigating the resource allocation problem. In particular, we first approximate the covariance matrix of the quantization error with a simple expression and derive an analytical expression of the received Signal-to-Noise Ratio (SNR) based on the deterministic equivalence theory. Then the three performance metrics (the spectral efficiency, energy efficiency, and total energy consumption) oriented problems are formulated analytically. With practical system requirements, these three metrics can be collaboratively optimized. Finally, we propose an optimization solver to derive the optimal partition of channel coherence time. Experiment results verify the benefits of the proposed resource allocation schemes under three different scenarios and illustrate the tradeoff of resource allocation between three stages.

The application of artificial intelligence technology in Internet of Vehicles (IoV) has attracted great research interests with the goal of enabling smart transportation and traffic management. Meanwhile, concerns have been raised over the security and privacy of the tons of traffic and vehicle data. In this regard, Federated Learning (FL) with privacy protection features is considered a highly promising solution. However, in the FL process, the server side may take advantage of its dominant role in model aggregation to steal sensitive information of users, while the client side may also upload malicious data to compromise the training of the global model. Most existing privacy-preserving FL schemes in IoV fail to deal with threats from both of these two sides at the same time. In this paper, we propose a Blockchain based Privacy-preserving Federated Learning scheme named BPFL, which uses blockchain as the underlying distributed framework of FL. We improve the Multi-Krum technology and combine it with the homomorphic encryption to achieve ciphertext-level model aggregation and model filtering, which can enable the verifiability of the local models while achieving privacy-preservation. Additionally, we develop a reputation-based incentive mechanism to encourage users in IoV to actively participate in the federated learning and to practice honesty. The security analysis and performance evaluations are conducted to show that the proposed scheme can meet the security requirements and improve the performance of the FL model.

Exploring open fields with coordinated unmanned vehicles is popular in academia and industries. One of the most impressive applicable approaches is the Internet of Vehicles (IoV). The IoV connects vehicles, road infrastructures, and communication facilities to provide solutions for exploration tasks. However, the coordination of acquiring information from multi-vehicles may risk data privacy. To this end, sharing high-quality experiences instead of raw data has become an urgent demand. This paper employs a Deep Reinforcement Learning (DRL) method to enable IoVs to generate training data with prioritized experience and states, which can support the IoV to explore the environment more efficiently. Moreover, a Federated Learning (FL) experience sharing model is established to guarantee the vehicles' privacy. The numerical results show that the proposed method presents a better successful sharing rate and a more stable convergence within the comparison of fundamental methods. The experiments also suggest that the proposed method could support agents without full information to achieve the tasks.

With the intelligentization of the Internet of Vehicles (IoVs), Artificial Intelligence (AI) technology is becoming more and more essential, especially deep learning. Federated Deep Learning (FDL) is a novel distributed machine learning technology and is able to address the challenges like data security, privacy risks, and huge communication overheads from big raw data sets. However, FDL can only guarantee data security and privacy among multiple clients during data training. If the data sets stored locally in clients are corrupted, including being tampered with and lost, the training results of the FDL in intelligent IoVs must be negatively affected. In this paper, we are the first to design a secure data auditing protocol to guarantee the integrity and availability of data sets in FDL-empowered IoVs. Specifically, the cuckoo filter and Reed-Solomon codes are utilized to guarantee error tolerance, including efficient corrupted data locating and recovery. In addition, a novel data structure, Skip Hash Table (SHT) is designed to optimize data dynamics. Finally, we illustrate the security of the scheme with the Computational Diffie-Hellman (CDH) assumption on bilinear groups. Sufficient theoretical analyses and performance evaluations demonstrate the security and efficiency of our scheme for data sets in FDL-empowered IoVs.

Although Federated Deep Learning (FDL) enables distributed machine learning in the Internet of Vehicles (IoV), it requires multiple clients to upload model parameters, thus still existing unavoidable communication overhead and data privacy risks. The recently proposed Swarm Learning (SL) provides a decentralized machine learning approach for unit edge computing and blockchain-based coordination. A Swarm-Federated Deep Learning framework in the IoV system (IoV-SFDL) that integrates SL into the FDL framework is proposed in this paper. The IoV-SFDL organizes vehicles to generate local SL models with adjacent vehicles based on the blockchain empowered SL, then aggregates the global FDL model among different SL groups with a credibility weights prediction algorithm. Extensive experimental results show that compared with the baseline frameworks, the proposed IoV-SFDL framework reduces the overhead of client-to-server communication by 16.72%, while the model performance improves by about 5.02% for the same training iterations.

The proliferation of internet communication channels has increased telecom fraud, causing billions of euros in losses for customers and the industry each year. Fraudsters constantly find new ways to engage in illegal activity on the network. To reduce these losses, a new fraud detection approach is required. Telecom fraud detection involves identifying a small number of fraudulent calls from a vast amount of call traffic. Developing an effective strategy to combat fraud has become challenging. Although much effort has been made to detect fraud, most existing methods are designed for batch processing, not real-time detection. To solve this problem, we propose an online fraud detection model using a Neural Factorization Autoencoder (NFA), which analyzes customer calling patterns to detect fraudulent calls. The model employs Neural Factorization Machines (NFM) and an Autoencoder (AE) to model calling patterns and a memory module to adapt to changing customer behaviour. We evaluate our approach on a large dataset of real-world call detail records and compare it with several state-of-the-art methods. Our results show that our approach outperforms the baselines, with an AUC of 91.06%, a TPR of 91.89%, an FPR of 14.76%, and an F1-score of 95.45%. These results demonstrate the effectiveness of our approach in detecting fraud in real-time and suggest that it can be a valuable tool for preventing fraud in telecommunications networks.

Due to their simple hardware, sensor nodes in IoT are vulnerable to attack, leading to data routing blockages or malicious tampering, which significantly disrupts secure data collection. An Intelligent Active Probing and Trace-back Scheme for IoT Anomaly Detection (APTAD) is proposed to collect integrated IoT data by recruiting Mobile Edge Users (MEUs). (a) An intelligent unsupervised learning approach is used to identify anomalous data from the collected data by MEUs and help to identify anomalous nodes. (b) Recruit MEUs to trace back and propose a series of trust calculation methods to determine the trust of nodes. (c) The last, the number of active detection packets and detection paths are designed, so as to accurately identify the trust of nodes in IoT at the minimum cost of the network. A large number of experimental results show that the recruiting cost and average anomaly detection time are reduced by 6.5 times and 34.33% respectively, while the accuracy of trust identification is improved by 20%.

Identifying rare patterns for medical diagnosis is a challenging task due to heterogeneity and the volume of data. Data summarization can create a concise version of the original data that can be used for effective diagnosis. In this paper, we propose an ensemble summarization method that combines clustering and sampling to create a summary of the original data to ensure the inclusion of rare patterns. To the best of our knowledge, there has been no such technique available to augment the performance of anomaly detection techniques and simultaneously increase the efficiency of medical diagnosis. The performance of popular anomaly detection algorithms increases significantly in terms of accuracy and computational complexity when the summaries are used. Therefore, the medical diagnosis becomes more effective, and our experimental results reflect that the combination of the proposed summarization scheme and all underlying algorithms used in this paper outperforms the most popular anomaly detection techniques.

A network intrusion detection system is critical for cyber security against illegitimate attacks. In terms of feature perspectives, network traffic may include a variety of elements such as attack reference, attack type, a sub-category of attack, host information, malicious scripts, etc. In terms of network perspectives, network traffic may contain an imbalanced number of harmful attacks when compared to normal traffic. It is challenging to identify a specific attack due to complex features and data imbalance issues. To address these issues, this paper proposes an Intrusion Detection System using transformer-based transfer learning for Imbalanced Network Traffic (IDS-INT). IDS-INT uses transformer-based transfer learning to learn feature interactions in both network feature representation and imbalanced data. First, detailed information about each type of attack is gathered from network interaction descriptions, which include network nodes, attack type, reference, host information, etc. Second, the transformer-based transfer learning approach is developed to learn detailed feature representation using their semantic anchors. Third, the Synthetic Minority Oversampling Technique (SMOTE) is implemented to balance abnormal traffic and detect minority attacks. Fourth, the Convolution Neural Network (CNN) model is designed to extract deep features from the balanced network traffic. Finally, the hybrid approach of the CNN-Long Short-Term Memory (CNN-LSTM) model is developed to detect different types of attacks from the deep features. Detailed experiments are conducted to test the proposed approach using three standard datasets, i.e., UNSW-NB15, CIC-IDS2017, and NSL-KDD. An explainable AI approach is implemented to interpret the proposed method and develop a trustable model.

A large number of network security breaches in IoT networks have demonstrated the unreliability of current Network Intrusion Detection Systems (NIDSs). Consequently, network interruptions and loss of sensitive data have occurred, which led to an active research area for improving NIDS technologies. In an analysis of related works, it was observed that most researchers aim to obtain better classification results by using a set of untried combinations of Feature Reduction (FR) and Machine Learning (ML) techniques on NIDS datasets. However, these datasets are different in feature sets, attack types, and network design. Therefore, this paper aims to discover whether these techniques can be generalised across various datasets. Six ML models are utilised: a Deep Feed Forward (DFF), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Decision Tree (DT), Logistic Regression (LR), and Naive Bayes (NB). The accuracy of three Feature Extraction (FE) algorithms is detected; Principal Component Analysis (PCA), Auto-encoder (AE), and Linear Discriminant Analysis (LDA), are evaluated using three benchmark datasets: UNSW-NB15, ToN-IoT and CSE-CIC-IDS2018. Although PCA and AE algorithms have been widely used, the determination of their optimal number of extracted dimensions has been overlooked. The results indicate that no clear FE method or ML model can achieve the best scores for all datasets. The optimal number of extracted dimensions has been identified for each dataset, and LDA degrades the performance of the ML models on two datasets. The variance is used to analyse the extracted dimensions of LDA and PCA. Finally, this paper concludes that the choice of datasets significantly alters the performance of the applied techniques. We believe that a universal (benchmark) feature set is needed to facilitate further advancement and progress of research in this field.

In this paper, we propose a doping approach to lower the error floor of Low-Density Parity-Check (LDPC) codes. The doping component is a short block code in which the information bits are selected from the coded bits of the dominant trapping sets of the LDPC code. Accordingly, an algorithm for selecting the information bits of the short code is proposed, and a specific two-stage decoding algorithm is presented. Simulation results demonstrate that the proposed doped LDPC code achieves up to 2.0 ​dB gain compared with the original LDPC code at a frame error rate of 10⁻⁶. Furthermore, the proposed design can lower the error floor of original LDPC codes.

In order to address the problems of the single encryption algorithm, such as low encryption efficiency and unreliable metadata for static data storage of big data platforms in the cloud computing environment, we propose a Hadoop based big data secure storage scheme. Firstly, in order to disperse the NameNode service from a single server to multiple servers, we combine HDFS federation and HDFS high-availability mechanisms, and use the Zookeeper distributed coordination mechanism to coordinate each node to achieve dual-channel storage. Then, we improve the ECC encryption algorithm for the encryption of ordinary data, and adopt a homomorphic encryption algorithm to encrypt data that needs to be calculated. To accelerate the encryption, we adopt the dual-thread encryption mode. Finally, the HDFS control module is designed to combine the encryption algorithm with the storage model. Experimental results show that the proposed solution solves the problem of a single point of failure of metadata, performs well in terms of metadata reliability, and can realize the fault tolerance of the server. The improved encryption algorithm integrates the dual-channel storage mode, and the encryption storage efficiency improves by 27.6% on average.