Accurate beam tracking in high-mobility vehicular networks has become a key issue for reliable communication link establishment. Recent work on integrated sensing and communication has shown the superiority of the sensing-assisted beam tracking. By estimating the angular parameters of vehicles based on the radar echoes, the roadside unit (RSU) can perform a so-called predictive beamforming to pair the beams with high accuracy as well as low latency. To further reduce the estimation error due to the nonlinear measurements, in this letter, we develop a deep learning-based approach to exploit the dependency between the angles and the measurements, and to obtain the angle estimates. Simulation results show that our proposed approach can improve the estimation performance and maintain reliable communications in high-mobility vehicular networks.

The digital twin (DT), which effectively represents the actual real-world physical system or process, has reshaped the classic manufacturing, construction, as well as healthcare industry. As for realizing DT, both sensing and communication functionalities are demanded, which fully builds the connectivity between the physical world and the digital world. We first conducted a survey on the current situation of DT combined with communication and sensing. Inspired from this survey and the current development of communication and sensing, in this paper, we attempt to study the communication annd sensing technologies of physical layer in DT, to reduce the hardware and spectrum overhead. First, we studied the degree of freedom (DoF) problem in general communication and sensing system, and contribute to the DoF definition in the sensing system. Then, in order to improve the spectrum efficiency in DT system, we proposed an iterative optimization framework to address the coexistence of communication and sensing, and some examples are provided. Finally, in order to pursue a better integration gain, we proposed a new waveform design method based on DoF completion. The proposed optimization method can achieve the mean square error (MSE) lower bound. Simulation results demonstrate the effectiveness of various problems in the above scenarios.

Beamforming improves the received signal power and eliminates undesirable interference by sharpening the transmitted signal toward a specific direction, enhancing service quality in the future vehicle network. However, the traditional beam codebook has gradually failed to cope with high-speed mobile services and complex pavement conditions due to beam misalignment and channel fading. To address the challenges above, this paper proposes a transformer-based beamforming approach to achieve sensing-assisted high reliable communication. We use the multimodal data collected by the sensors at the base station for beamforming to optimize the communication performance. The proposed model employs three-dimensional (3D) ResNet-18 to extract multimodal features and leverages the transformer's merged-attention mechanism to fuse these features for beamforming. The experimental result based on real-world vision, radar, LiDAR, and position data shows the advance of our proposed method, which achieves 91.59% top-3 accuracy on average and exceeds over 30% top-1 accuracy than single-modal schemes in the high-speed environment.

Deep learning (DL) has been widely applied in automatic modulation classification (AMC), while the superb performance highly depends on high-quality datasets. Motivated by this, the AMC under few-shot conditions is considered in this letter, where a novel network architecture is proposed, namely automatic modulation classification relation network (AMCRN), and verified with the baseline methods. Experimental results state that the accuracy of proposed AMCRN exceeds 90% and 10% to 50% improvements are obtained compared with classical schemes when the signal-to-noise ratio (SNR) is greater than -2 dB.

This paper develops a new approach of large language model (LLM) enhanced reconfigurable intelligent surface (RIS), in a bid to achieve energy-efficient and reliable communication in 6G Internet of Vehicles (IoV). It is well known that RIS offers an innovative solution to improve signal quality by intelligently adjusting the propagation path of radio waves. However, configuring RIS in the dynamically changing vehicular environment remains a challenge. This study leverages the analytical capabilities of LLM, combined with key IoV data such as channel status, vehicle movement patterns, and quality of service requirements, to model and optimize RIS-based IoV communication systems. The work focuses on constructing a realtime model of the RIS-based IoV wireless transmission system and proposes an optimal strategy for wireless resource allocation. Through comprehensive simulation results, this paper justifies the significant performance advantages of LLM-enhanced RIS, demonstrating a viable technical pathway for the development of 6G IoV.

Federated learning (FL) enables collaborative training of machine learning models across distributed medical data sources without compromising privacy. However, applying FL to medical image analysis presents challenges like high communication overhead and data heterogeneity. This paper proposes novel FL techniques using explainable artificial intelligence (XAI) for efficient, accurate, and trustworthy analysis. A heterogeneity-aware causal learning approach selectively sparsifies model weights based on their causal contributions, significantly reducing communication requirements while retaining performance and improving interpretability. Furthermore, blockchain provides decentralized quality assessment of client datasets. The assessment scores adjust aggregation weights so higher-quality data has more influence during training, improving model generalization. Comprehensive experiments show our XAI-integrated FL framework enhances efficiency, accuracy and interpretability. The causal learning method decreases communication overhead while maintaining segmentation accuracy. The blockchain-based data valuation mitigates issues from low-quality local datasets. Our framework provides essential model explanations and trust mechanisms, making FL viable for clinical adoption in medical image analysis.