New radio (NR) and non-orthogonal multiple access (NOMA) offer scalable and efficient resource allocation in Beyond 5G (B5G) networks. NR implements mixed numerology with flexible frame structures for future compatibility, whereas NOMA allows users with different channel states to share an identical Physical Resource Block (PRB). Multi-connectivity enables a user to connect to multiple networks for reliability, and multicast conveys data to users simultaneously that request the same content. However, resource allocation in the NOMA-based mixed numerology system with multi-connectivity for multicast remains unexplored. The problem is challenging due to 1) the different shapes of PRBs in NR and 2) the shared locations of PRBs in a frame with NOMA. In this paper, we formulate a new optimization problem, named Multicast, Multi-connectivity, and Multi-Dimensional Resource Allocation Problem (M3DRAP), and prove its NP-hardness and inapproximability. We propose an approximation algorithm for general M3DRAP with the ideas ofMulticast Inter-Numerology Relation,Layer Dissimilarity,Subgrouping Nonuniformity, andSegmentation Preference. To find the intrinsic properties of PRB allocation for multicast in NOMA-based networks, we consider a single B5G usage scenario (e.g., eMBB, URLLC, or mMTC) and propose another approximation algorithm. Simulations demonstrate our algorithms improve the weighted sum rate by over 50% and increase the user satisfaction ratio by 1.5x.

With the emergence of new applications for holographic-type communication in healthcare, entertainment, and education, point cloud video transmission has become essen-tial. This paper aims to reduce the bandwidth cost by leveraging tile-based video transmission with point cloud registration in a wireless multicast network. A point cloud video is divided into multiple tiles, and each tile contains a portion of point cloud objects and can be registered by adjacent tiles with some similar objects under the registration rotation and registration overlap constraints. We formulate a new optimization problem and prove that it is NP-hard, and then we design an algorithm Multicast Multi-Tile Registration (MMTR) to select multicasting and registered tiles under consideration of socially related users' preferences with the idea of a tile registration graph. A more popular tile can be multicasted to more friends to minimize the bandwidth cost. Experimental results with real datasets show that MMTR can reduce bandwidth costs by more than 20% and achieve better video quality compared to state-of-the-art point cloud transmission algorithms.

By integrating the data from sensors and cameras, Digital Twin (DT) creates a virtual representation of real-world road traffic and road users (RUs) to enhance understanding and decision-making for traffic applications. However, modern traffic environments require sensors on RUs to create a virtual representation. Without these sensors, it is difficult to create virtual representations for RUs in the DT, limiting the system’s effectiveness. In this paper, we first design a DT-based system for pairing and updating the information of RUs in road traffic environments to support for RUs with and without sensors. Our system includes two modules for updating the information: 1) Localization module for updating the location of the RU itself, and 2) Road User detection module to detect the other RUs for updating. Then, we formulate a new optimization problem to minimize the Age of Incorrect Information (AoII) metric and propose an algorithm, named AoII minimization by Update Selection in Digital Twin (AoII-USDT), to determine the updating policy of each RU. Simulation results show that AoII-USDT outperforms state-of-the-art algorithms regarding total AoII, freshness improvement, and accuracy.

The emergence of AI has shifted the focus of wireless communications towards deep semantic-level fidelity (i.e., semantic communication networks), emphasizing the significance and effectiveness of transmitted data. However, semantic feature selection considering multicast users with social relations for feature sharing has not been explored. In this paper, we formulate a new optimization problem to minimize the total communication, forwarding, and computation costs, with the proof of NP-hardness and inapproximability. We propose a new algorithm, Multicast Semantic Feature Selection (MSFS), with the notions of Cross Task Semantic Indicator, Substituted Subgraph, and Socially Feature Selection Indicator, to select features on different mobile edge computing servers and cluster the users to receive features via multicast. Simulations with real datasets manifest that MSFS can reduce the total cost by more than 50% compared with state-of-the-art algorithms.

The localization of Road Users (RUs) is an important issue in adverse event prevention due to the unreliable nature of GPS and the high cost of high-precision location acquisition sensors. In addition, previous research on adverse event prevention on roads has not taken into account RUs in blind spots at the same time. To address these issues, we investigate a collaboration system for heterogeneous RUs and Mobile Edge Computing (MEC) servers, called Collaborative Adverse Event Prevention system (CAEP) to efficiently alert RUs to potential adverse events and perceive the blind spot of the RUs. CAEP includes two AI-based functional modules, a localization module and a blind spot detection module, and an adverse event prevention algorithm. The localization module localizes each RU and the blind spot detection module detects the other RUs in the blind spot. The adverse event prevention algorithm jointly considers general road collision events and the event of a difference in radius between the inner wheels of a vehicle to completely include adverse events on the road. We implement CAEP in a real-world traffic environment with Jetson AGX Xavier devices and cameras to evaluate the performance. Our evaluation shows that CAEP provides RUs with sufficient preparation time to prevent adverse events and correctly detects the RUs in blind spots.

New radio (NR) and non-orthogonal multiple access (NOMA) have emerged for more scalable and efficient resource utilization in 5G. NR implements mixed numerology with a flexible radio frame structure to ensure forward compatibility for future services, whereas NOMA allows multiple users with different channel states to share identical radio resources. However, the resource allocation in the NOMA-based mixed numerology system is challenging due to the naturally different shapes of Physical Resource Block (PRB) for NR and the reused locations of PRBs in a radio frame for NOMA. In this paper, we formulate a new optimization problem Multi-Dimensional Resource Allocation Problem (MDRAP) and prove that MDRAP is NP-hard. To solve the problem, we propose an approximation algorithm to maximize the weighted sum rate under the heterogeneity of users. The algorithm includes Zone Displacement to displace the locations of allocated PRBs in different layers of the radio frame, and Zone Allocation to change the location of the bounded rectangles (i.e., zones) for the allocation in each layer. We design Layer Dissimilarity to examine the location and shape of PRBs for avoiding inter-numerology interference between different layers. Simulation results show that the proposed algorithm outperforms state-of-the-art algorithms regarding throughput and fairness.