Haixia Zhang’s research while affiliated with Shandong University and other places

Caching 360-degree videos at the network edge can reduce user content request latency and mitigate transmission congestion in backbone networks. Given the fact that user only views a part of content of 360-degree scope at any time, caching the entire video is resource inefficient. To address this, we focus a Multi-access Edge Computing (MEC)-based 360-degree video service system, where the edge server only caches a portion of each video that is most likely falling in the Field of View (FoV) of users. To minimize the average video request latency of all users in the system, we formulate a large-scale 0-1 knapsack problem, which is NP-hard. To tackle it, we proposed a heuristic algorithm where the user viewing patterns extracted from the historical request information are taken into account. Specifically, we first design a cascading cache space allocation method to assign the total cache space of edge server to each segment of videos. After that, the original problem is decomposed into several small-scale yet individual tile caching subproblems with compressed solution space. Then, they are solved by using the dynamic programming algorithm with moderate complexity. To further enhance the caching performance, the PSO-based algorithm is designed to fine tune the parameters involved in the proposed caching algorithm. In addition, we introduce a content-based method to calculate the request probability of the newly generated videos. The effectiveness of the proposed algorithm is evaluated through simulations based on a real world dataset, where the results demonstrate a substantial improvement in both video request latency and cache hit rate compared to the benchmark methods.

Vehicle-to-everything (V2X) underlaid cellular networks in underlaid mode suffers catastrophic co-channel interference caused by spectrum sharing, results in a reduced system sum-rate. To cope with this, this work studies a social-mobility-aware non-orthogonal multiple access (NOMA)-enabled V2X underlaid cellular network to mitigate the co-channel interference and improve the sum-rate. By jointly optimizing vehicle pairing and resources, a sum-rate maximization problem is formulated under the diverse quality of service requirements of both cellular and vehicular users. The formulated problem is proved to be a non-deterministic polynomial-time (NP)-hard problem, and is difficult to solve. As an alternative, we propose a NOMA-based joint vehicle pairing, spectrum assignment and power control algorithm (NOMA-JVP-SA-PCA), with which the original problem is decomposed into two disjoint subproblems, i. e., 1) joint vehicle pairing and spectrum assignment subproblem, and 2) power control subproblem. Dealing the first subproblem, we propose a heuristic social-mobility-aware vehicle pairing algorithm (HSMA-VPA) and a revised Kuhn-munkres-based spectrum assignment algorithm (KM-SAA) to acquire the vehicle pairing and spectrum assignment solutions. Then, solving the second subproblem, a closed-form power solution is obtained utilizing a three-dimensional geometric power control approach (3D-PCA). Finally, we solve the original problem through an iterative method. Simulation results show that the proposed NOMA-JVP-SA-PCA effectively enhances the sum-rate and outperforms the baseline algorithms around 24%-53% within a specific range.

Reconfigurable intelligent surface (RIS) is widely recognized as one of the key technologies for 6G due to its ability to enhance communication signal coverage and quality. To fully explore the potential of RIS, its element phase shift must be optimized. Traditional schemes dedicated to the parameter design of RIS, such as finite element methods (FEM) and ray-tracing, are not suitable for handling dynamic scenarios due to their high computational complexity. Therefore, the digital twin has been adopted, but it is still far from working in real time. To address this, we introduce and combine physics informed neural networks (PINN) with digital twin to build digital replicas of RIS in virtual space for dynamic channel conditions. The proposed physics informed digital twin architecture integrates physical and data information, consisting of sensing, modeling, real-time interaction, prediction, and phase shift optimi-zation. In this work, the details on how to integrate PINN with digital twin to model RIS-assisted wireless communications are described, and a use case performance is analyzed. In addition, possible research directions and challenges are discussed. Simulation results show that the proposed scheme greatly reduces the prediction time compared to FEM and achieves better accuracy.

This chapter focuses on the secure transmission of wireless-powered full-duplex (FD) relay systems, where a multiple-antenna source communicates with a single-antenna destination with the help of an FD relay in the presence of a single-antenna eavesdropper. It is assumed that the FD relay is wirelessly energy harvesting-enabled, adopting both transmit and receive antennas to harvest energy in time switching (TS) mode. As the objective of this chapter is to maximize the system secrecy rate by jointly designing the energy beamforming vector, the information beamforming vector, and the TS coefficient, an optimization problem is formulated in Sect. 6.2. The formulated problem is proven to be nonconvex, and the challenge is to concurrently solve out the three variables. To address this difficulty, an iterative algorithm is proposed in Sect. 6.3 to convert the formulated optimization problem into three convex subproblems, on which the closed-form solutions for the beamforming vectors are derived and the TS coefficient is obtained. The convergence property of the iterative method is analyzed. Simulations are performed in Sect. 6.4 to verify the theoretical derivations in terms of the convergence speed and the secrecy rate. The results reveal that the secrecy rate performance of exploiting the transmit antenna together with the receive antenna for energy harvesting at the FD relay outperforms that of only the receive antenna case. Moreover, although loopback interference exists between the antennas, FD relaying can substantially increase the secrecy rate compared with the half-duplex relaying architecture. Section 6.5 concludes the work.

In this chapter, we summarize the main results and contributions of this study in Sect. 7.1. Numerous relevant future research directions are presented in Sect. 7.2, including context-aware resource allocation, autonomous resource management, edge computing and edge intelligence, AI for resource optimization, federated learning for resource optimization, privacy-preserving resource management, interoperability and standardization, simulation, and testbed development.

This chapter investigates the socially aware proactive edge caching strategy in vehicular social networks, where vehicles can be selected as caching nodes to assist content delivery. The objective is to achieve a trade-off between the cost of providing caching services and the content transmission latency. This strategy presents two challenges: (1) which vehicles can be selected as caching nodes, and (2) how to place content on these selected vehicles without violating user privacy. To address these issues, a novel community detection and attention-weighted federated learning-based proactive edge caching (CAFLPC) strategy is proposed. In this strategy, we first group vehicles into different communities on the basis of both the mobility and social properties of the vehicles and then select important vehicles (IVs) as caching nodes for each community by considering the social importance of the vehicles. To determine how to place the popular content in these selected IVs, an attention-weighted federated learning (AWFL)-based content popularity prediction framework is proposed. It integrates attention-weighted federated learning with a bidirectional long short-term memory network (AWFL_BiLSTM) to achieve higher content popularity prediction accuracy while protecting user privacy. Considering imbalances in the active levels and local computing capacities of the vehicles, an attention-weighted aggregation mechanism is proposed to improve training efficiency and prediction accuracy. The simulation results show that the proposed CAFLPC strategy outperforms existing caching strategies by approximately 2.2~35.1% in terms of transmission latency, which is reduced per unit cost.

Existing computation and communication (2C) optimization schemes for cellular-based vehicle-to-everything (C-V2X) networks do not consider the influence of social trust. Computational tasks may be offloaded to untrusted vehicles, hindering the accurate execution of computational tasks. This may lead to a re-offloading of the computational tasks, consuming additional power, and decreasing the energy efficiency (EE) for offloading. To address this issue, this work devotes itself to investigating the social-mobility-aware underlying C-V2X framework and proposes a novel EE-oriented 2C assignment scheme. In doing so, we assume that the task vehicular user (T-VU) can offload computational tasks to the service vehicular user (S-VU) and the road side unit (RSU). In Sect. 5.2, an EE maximization problem to assign 2C resources simultaneously through joint optimization is formulated, which is a mixed-integer nonlinear programming (MINLP) problem. To solve this problem, we transform it into separate computation and communication resource allocation subproblems in Sect. 5.3. To address the first subproblem, we fully integrate the social and mobility characteristics and design a heuristic algorithm to achieve edge server selection and task splitting. To address the complex co-channel interference in the second subproblem, the power allocation and spectrum assignment solutions are obtained via a tightening lower bound method and a Kuhn-Munkres (KM) algorithm. Finally, we solve the original problem through an iterative method. In Sect. 5.4, the simulation results show that the proposed scheme can significantly enhance the system EE. Section 5.5 concludes the work.

To support the ever-expanding demands for multifarious vehicular services with a limited spectrum, vehicle-to-everything (V2X) networks, which underlay cellular networks, have drawn extensive attention. The underlaid mode suffers from catastrophic cochannel interference caused by spectrum sharing between cellular users and V2X users, thus reducing the sum rate of the system. The existing solutions for mitigating cochannel interference focus mainly on the physical domain without considering the influence of the social domain. This may greatly limit the sum-rate enhancement potential of the utilized system. To address this issue, a social mobility-aware V2X underlaying a cellular network is studied in this chapter. By jointly optimizing the vehicle pairing situation and resources (i.e., the spectrum and power), a sum-rate maximization problem is formulated for V2X-underlaid cellular networks while satisfying the diverse quality of service (QoS) requirements of both cellular users and vehicular users. The formulated problem is proven to be a nondeterministic polynomial-time (NP)-hard problem and is difficult to directly solve. As an alternative, we propose a joint vehicle pairing, spectrum assignment, and power control algorithm in Sect. 3.3. In this section, the original problem is decomposed into two disjoint subproblems, i.e., (1) a joint vehicle pairing and spectrum assignment subproblem and (2) a power control subproblem. To address the first subproblem, we propose a heuristic social mobility-aware vehicle pairing algorithm (HSMA-VPA) and a revised Kuhn-Meyer-based spectrum assignment algorithm (KM-SAA) to acquire the vehicle pairing and spectrum assignment solutions. Then, by solving the second subproblem, a closed-form power solution is obtained via a three-dimensional geometric power control approach (3D-PCA). Finally, we solve the original problem through an iterative method. Simulation results show that the proposed NOMA-JVP-SA-PCA effectively enhances the sum rate and outperforms the baseline algorithms by approximately 24–53% in Sect. 3.4.

To help readers better understand the motivation for applying different technologies to vehicular social networks (VSNs), a comprehensive overview of current VSNs is presented in this chapter, including the architecture, characteristics, classifications, applications, and importance of VSNs in modern transportation systems. Then, the challenges facing VSNs in terms of resource allocation and vehicle behavior prediction are detailed. Next, the resource management problems for VSNs and methods for solving them, including strategies for communication resource management, computational resource allocation, caching resource management, and multidimensional resource allocation, are surveyed. Finally, the key research problems investigated in this monograph are presented.

Wireless traffic prediction has drawn increasing research interest because it can provide network optimization guidance. With the predicted information, one can preassign resources on demand and adaptively perform network congestion control. The efficiency of the utilized network is therefore enhanced. However, conducting wireless traffic prediction in the context of mobile scenarios, such as the Internet of Vehicles (IoV), is still challenging. The mobile nature of vehicles, which dynamically changes the topology of the constructed network, makes prediction difficult. This chapter focuses on implementing deep learning-based wireless traffic prediction in the IoV scenario. Section 2.1 proposes a novel method for matching the movement and communication behaviors of vehicles by merging two independent datasets containing the trajectories of vehicles and communication traffic volumes. Then, a novel STeP-UNet is proposed in Sect. 2.2, in which a spatiotemporal partial (STeP) convolutional neural network module is embedded to capture the cross-domain features of the observed wireless traffic pattern, and the UNet structure is utilized to realize skip connections from the front layer to the back layer to fuse different resolutions. The experimental results confirm the promising performance of the proposed model in Sect. 2.3, where a 4~8% performance improvement can be achieved over other benchmark methods.