Frequency hopping-based anti-jamming techniques are not always applicable in vehicular ad-hoc networks (VANETs) due to the high mobility of onboard units (OBUs) and the large-scale network topology. In this paper, we use unmanned aerial vehicles (UAVs) to relay the message of an OBU and improve the communication performance of VANETs against smart jammers that observe the ongoing OBU and UAV communication status and even induce the UAV to use a specific relay strategy and then attack it accordingly. More specifically, the UAV relays the OBU message to another roadside unit (RSU) with a better radio transmission condition if the serving RSU is heavily jammed or interfered. The interactions between a UAV and a smart jammer are formulated as an anti-jamming UAV relay game, in which the UAV decides whether or not to relay the OBU message to another RSU, and the jammer observes the UAV and VANET strategy and chooses the jamming power accordingly. The Nash equilibria of the UAV relay game are derived to reveal how the optimal UAV relay strategy depends on the transmit cost and the UAV channel model. A hotbooting policy hill climbing (PHC)-based UAV relay strategy is proposed to help VANET resist jamming in the dynamic game without being aware of the VANET model and the jamming model. Simulation results show that the proposed relay strategy can efficiently reduce the bit error rate of the OBU message and thus increase the utility of the VANET compared with a Q-learning based scheme.

Recently, the development of the Metaverse has become a frontier spotlight, which is an important demonstration of the integration innovation of advanced technologies in the Internet. Moreover, artificial intelligence (AI) and 6G communications will be widely used in our daily lives. However, the effective interactions with the representations of multimodal data among users via 6G communications is the main challenge in the Metaverse. In this work, we introduce an intelligent cross-modal graph semantic communication approach based on generative AI and 3-dimensional (3D) point clouds to improve the diversity of multimodal representations in the Metaverse. Using a graph neural network, multimodal data can be recorded by key semantic features related to the real scenarios. Then, we compress the semantic features using a graph transformer encoder at the transmitter, which can extract the semantic representations through the cross-modal attention mechanisms. Next, we leverage a graph semantic validation mechanism to guarantee the exactness of the overall data at the receiver. Furthermore, we adopt generative AI to regenerate multimodal data in virtual scenarios. Simultaneously, a novel 3D generative reconstruction network is constructed from the 3D point clouds, which can transfer the data from images to 3D models, and we infer the multimodal data into the 3D models to increase realism in virtual scenarios. Finally, the experiment results demonstrate that cross-modal graph semantic communication, assisted by generative AI, has substantial potential for enhancing user interactions in the 6G communications and Metaverse.

With the rapid advancement of artificial intelligence (AI) technology, the Internet of Vehicles (IoV) is becoming increasingly important in intelligent transportation systems (ITSs). At the same time, large language models (LLMs) and generative AI (GenAI) are gradually playing significant roles in the IoV. Multimodal LLMs (MLLMs) are models capable of integrating multiple modalities of information, enhancing the environmental perception capabilities of the IoV. GenAI can generate highly complex virtual driving scenarios, which are used to test and optimize intelligent driving algorithms, reducing the risks and costs of real-world road testing. Embodied AI can interact with the environment and make decisions, and can learn in the simulated environments that are provided by GenAI. To address issues like low correlation of multimodal data, high-uncertainty of driving environment, and insufficient intelligence in the IoV, this article proposes a perception and decision-making system empowered by embodied AI. First, we utilize a transformer to achieve multimodal data fusion integrated with vehicle-to-vehicle (V2V) data. Second, the MLLM analyzes the situation based on driving intentions, communication conditions, and the perceived environmental expressions. Finally, the MLLM makes the optimal decisions and adjustments based on the analysis. The experimental results show that the integration of V2V enables the IoV to more comprehensively perceive a complex driving environment. The MLLM can empower the IoV to make accurate decisions in high-uncertainty scenarios, significantly improving route completion rates.

Cross-modal communications that incorporate audio-visual and tactile signals will bring a more holistic immersive experience to people. However, due to the different transmission requirements of these signals, it is a challenging task to rationalize the allocation of transmission resources. Therefore, this work proposes a joint transmission scheme to deal with the resource allocation problem of diverse signals. Network slicing and puncturing architecture are introduced in the scheme to achieve flexible resource allocation and reduce the wasting of resources. To reduce the negative impact of puncturing transmission on users, we construct the optimization problem related to transmission rate and reliability. This problem can realize the reasonable allocation of radio resources and meet the transmission requirements of the two types of signals. Next, we divide the optimization problem into two parts: video traffic resources allocation and tactile traffic puncturing resources allocation. To solve both of the problems, we leverage the channel matching (CM) algorithm and puncturing resource allocation (PRA) algorithm. In addition, we discuss the advantages and disadvantages of both ways to occupy puncturing resources, namely, occupy resources proportionally (ORP) and occupy resources blocks for transmission (ORB). Finally, the effectiveness of the proposed scheme is verified by comparing the excepted rate and resources loss ratio of system users with different schemes.

The evolution from fifth-generation mobile communications (5G) to beyond 5G (B5G) will lead to more ubiquitous and smarter paradigms in the Internet of Things (IoT). Communication will shift from the classical information theory to a semantic communication (SemCom) paradigm driven by artificial intelligence (AI) to enhance the capacity and optimize resources. Image SemCom (ISC) will empower IoT applications, such as drone image acquisition. However, ISC requires suitable devices and sufficient computing resources to support complex neural network models, posing a significant challenge. To address this, we propose a federated semantic feature distillation (FedSFD) architecture to improve the global performance of ISC by combining federated learning (FL) and feature distillation (FD) for the feature knowledge transfer. First, lightweight IoT device models in the edge group and the powerful server model alternately minimize losses to update parameters. After learning middle-layer features of all the edge models, the server can guide the individual device models. Second, we incorporate the information bottleneck (IB) concept into the design of loss functions to balance compression and reconstruction. Third, focusing on the tradeoff between the local training and knowledge interaction, FedSFD achieves image semantic reconstruction without sharing the private data, ensuring personalization and privacy protection in the FL framework. Finally, compared to the baseline, the simulation experiments show that the proposed approach achieves better ISC reconstruction and noise robustness during group ISC.

Semantic communication represents a paradigm shift in 6G communications, emphasizing the transmission of meaning rather than solely syntactic elements. Despite this advancement, current approaches exhibit limitations in generality, adaptability to dynamic environments, and dependence on static knowledge bases. To address these challenges, we propose a novel Large Language Model-based Semantic Communication (LLM-SemCom). LLM-SemCom incorporates three key innovations: 1) structured semantic triple representation that mitigates LLM hallucinations unlike existing unstructured approaches, 2) knowledge-base-free LLM semantic processing that adapts dynamically without static domain constraints, and 3) Retrieval-Augmented Generation-enhanced personalization that maintains semantic fidelity while enabling user-specific adaptation. Experimental results demonstrate that LLM-SemCom significantly outperforms existing methods, achieving up to a 22.7% improvement in sentence similarity while maintaining consistent performance across diverse languages and channel conditions.