In autonomous driving scenarios, limited sensing range of individual autonomous vehicles (AVs) and exponential growth of sensing data have drawn increasing attention. This paper focuses on an integrated system combining communication and computation assistance for sensing enhancement, exploring functional fusion and performance optimization of the autonomous driving integrated sensing, communication, and computation (ISCC) system. Specifically, we first establish a cloud-edge-terminal collaborative ISCC system tailored for autonomous driving. For this system, we model sensing, communication, and computation separately, where the sensing model incorporates task-dependent characteristics, specifically considering the sequential execution of detection and tracking as well as the parallel nature of localization. Given that the collaborative computation between AVs and edge nodes aims to maximize system performance, we formulate a mixed integer nonlinear optimization problem. To solve this problem, we design two independent agents for resource and offloading configuration based on deep reinforcement learning. The former can adaptively allocate resources in each time slot without requiring prior knowledge of task arrival times, while the latter employs a partial offloading strategy to leverage the local computing capabilities of AVs, thereby addressing the limitations of existing approaches that rely on fixed resource allocation or neglect local computation. The simulation results show that the average task completion rate of the proposed scheme is significantly improved, the system cost is notably reduced compared with traditional schemes.

The Internet of Autonomous Vehicles (IoAV) faces growing challenges in user privacy and communication security, stemming from dynamic network topologies induced by highspeed vehicle mobility, resource-constrained onboard devices, and the inherent tension between identity anonymity and traceability in latency-critical applications. Given the distributed architecture of fog computing and the limited storage and computational capabilities of vehicles, conventional anonymous authentication and centralized key negotiation mechanisms prove insufficient in addressing these issues. In response, We propose a distributed authentication and key negotiation protocol that combines multifactor biometrics, zero-knowledge proof (ZKP), and physical unclonable function (PUF) without relying on a trusted third party. Specifically, we design an efficient ZKP algorithm based on Chebyshev polynomials with low overhead and strong anonymity. Our key innovation is the implementation of independent key negotiation of three untrusted entities in a single protocol cycle, enabling 23 security features and functions. The performance analysis shows that the scheme takes only 17 ms to complete the protocol flow, and it reduces vehicle memory usage by 33% to 83%, service latency by 61% to 83%, and communication overhead by 12% to 50% compared to existing schemes.

As urban rail transit evolves, the convergence of digitalization, networking and intelligence has emerged as a pivotal trend, accompanied by the surge of intensive computing tasks and real-time demand. Edge computing is a promising solution to address the local resource constraints of various application devices covered by various subsystems within Urban Rail Transit Systems (URTS). In this paper, a Digital Twin (DT) assisted 5G-enabled edge computing network in URTS is established. We consider the heterogeneous service requirements of Ultra-Reliable Low-Latency Communication (URLLC) and Enhanced Mobile Broadband (eMBB) in different intelligent applications, and accordingly, we develop a task execution queue model. Additionally, to improve the task processing performance of the system, a task offloading and resource allocation scheme based on the Multi-Agent Twin Delayed Deep Deterministic policy gradient (MATD3) algorithm is proposed. User Equipment (UEs) and Edge Servers (ESs) are treated as distinct agents, accounting for their collaborative computation capabilities and differences in decision-making. A framework is established for centralized training in a DT assisted system with decentralized execution by each agent. Simulation results demonstrate that the proposed approach ensures the stringent latency requirement for URLLC tasks, enhances the Transaction Per Time Slot (TPTS) for eMBB tasks, and significantly reduces processing delays across all tasks. In addition, the proposed scheme also achieves load balance between UEs and ESs.

Although vehicular ad hoc networks (VANETs) significantly enhance traffic convenience, the propagation of erroneous information by malicious vehicles remains a challenging issue. To maintain message reliability, it is crucial to establish a trust management model that can promptly detect malicious vehicles and identify false messages. This article presents a novel trust management model based on blockchain, machine learning, and active detection technology. In the proposed model, we designed a trust evaluation scheme to evaluate the credibility by calculating the direct and indirect trust of the vehicle. To achieve this goal, we use active detection technology to detect indirect trust in vehicles, and then store it in the blockchain. The direct trust of the vehicle is calculated using a Bayesian classifier. The use of active detection technology speeds up the process of filtering out malicious vehicles. Machine learning technology simplifies the complex iterations involved in computing the trust value. Finally, the use of blockchain ensures the consistency and tamper-proofing of the trusted data. The simulation outcomes demonstrate that our approach outperforms the present trust management models.

Traffic flow prediction is a difficult undertaking in transportation systems, due to the intricate periodicity and real-time dynamics for traffic data, spatial-temporal dependency for road networks, existing prediction approaches fail to yield satisfactory results. We propose a traffic flow prediction method named Extended Multi-component External Interactive Gated Recurrent Graph Convolutional Network (EMGRGCN). The extended multi-component (EMC) module is incorporated into the prediction model to address the periodic temporal diffusion problem. Then, we introduce an encoder-decoder architecture that incorporates attention mechanism to capture spatial-temporal dependencies. Specifically, an External Interactive Gated Recurrent Unit (EIGRU) is utilized to capture crucial temporal features. EIGRU and graph convolutional network are combined in the encoder to extract spatial-temporal correlation, and EIGRU and convolutional neural network based decoder transforms the spatial-temporal characteristics into a sequence to predict future traffic flows. Experiments on public transportation datasets PEMSD8 and PEMSD4 demonstrate that EMGRGCN model achieves the best performance.

An end-to-end learning framework is proposed to optimize each module jointly in the communication system. Recently, convolutional neural network (CNN) and conditional Generative Adversarial Network (cGAN) are used for end-to-end learning. However, deeper network layers will degrade the effect of CNN. cGAN suffers from unstable training and lacks generative diversity. In this article, we propose the end-to-end learning based on deep residual network (ResNet) and Wasserstein GAN (WGAN) for communication with unknown channels (ResNet-WGAN). First, ResNet is applied to solve the problem of network degradation to extract deeper data features. Second, for unknown channels, WGAN with conditional information is used to fit the channel effect to improve training stability and generative diversity. Finally, we present the simulation results of the ResNet-WGAN under additive white Gaussian noise (AWGN) channel, Rayleigh fading channel, and frequency selective channel. The results demonstrate that the ResNet-WGAN reduces the communication bit error rate (BER) and block error rate (BLER). In particular, this article applies ResNet-WGAN to the Internet of Vehicles (IoV) communication, and the results demonstrate that ResNet-WGAN is more effective.

As the potential candidate of next generation communication networks, rate splitting multiple access (RSMA) has drawn great attentions. Usually to maximize the sum rate of RSMA, weighted minimum mean square error (WMMSE) algorithm is often used. But its computational complexity hinders its practical application. In this paper, we apply a deep learning network called Deep Unfolding (DU) to RSMA in Satellite-Terrestrial Integrated Network (STIN) with the intention of boosting downlink capacity. The momentum accelerated projection gradient descent (PGD) algorithm is adopted to substitute the complex operations and speed up the computations. By selecting the momentum and step size as the trainable parameters, the simulation results indicate that the deep learning approach outperforms the original WMMSE algorithm in sum rate and speed.

Channel estimation in time division duplexing (TDD)‐based massive multiple‐input multiple‐output (MIMO) systems is heavily hampered by the pilot contamination, which constitutes a major bottleneck on the overall system performance. This study considers the pilot contamination problem in multi‐cell TDD‐based massive MIMO systems, and analytical expressions are presented on the normalised mean square error (NMSE) of the minimum mean square error channel estimation algorithm. Based on the obtained NMSE, this study proposes an optimal pilot assignment strategy to minimise the effect of pilot contamination. In order to further improve the system performance, a pilot design‐based channel estimation scheme is proposed, where Chu sequences with perfect auto‐correlation property are employed to design the optimal pilot sequences aiming at acquiring the accurate channel state information. Simulation results show that the proposed pilot assignment strategy outperforms the random pilot assignment method, and approaches to the performance of the exhaustive search method which requires high computational complexity. Moreover, the performance gain of the pilot design‐based channel estimation scheme is verified in massive MIMO systems.