Digital-Twin-Assisted Safety Control for Connected Automated Vehicles in Mixed-Autonomy Traffic

The Butterfly Effect in Vehicular Digital Twin Systems: Complexity and Risk Analysis for Mixed-Traffic Scenarios

A Cloud–Edge Collaborative Architecture for Multimodal LLM-Based Advanced Driver Assistance Systems in IoT Networks

Toward High-Accuracy and Low-Latency Group Vehicle Trajectory Prediction with Linear UNet-Enhanced Fully Connected Spatial-Temporal Graph Neural Network

Split Learning with Differential Privacy for Integrated Terrestrial and Non-Terrestrial Networks

Intelligent Parking Service System Design Based on Digital Twin for Old Residential Areas

Due to the increasing number of vehicles and the limited land supply, old residential areas generally face parking difficulties. An intelligent parking service is a critical study direction to address parking difficulty since it can achieve the automatic management of parking processes and planning of parking spaces. However, the existing intelligent parking service systems have shortcomings such as low information quality, low management efficiency, and single service mode. To address the shortcomings, in this paper, we conduct a systematic study on utilizing digital twin (DT) technology to improve the intelligent parking service system. The main contributions are threefold: (1) We analyze the function requirements of the intelligent parking service for old residential areas, such as visual monitoring, refined management, and simulation optimization. (2) We design a DT-based intelligent parking service system by collecting data on physical parking space, constructing the corresponding virtual parking space, and building the user interaction platform. An old residential area in Guangzhou, China is used as a use case to show that the designed parking service system can meet the function requirements. (3) Through mathematical modeling and simulation evaluation, we utilize two typical intelligent parking services including dynamic parking planning and driving safety assessment to demonstrate the effectiveness of the proposed system. This study provides innovative solutions for parking management in old residential areas, utilizing DT technology to not only improve information quality and management efficiency, but also provide a theoretical basis and practical reference for the intelligent transformation of urban parking services.

D-Tracking: Digital Twin Enabled Trajectory Tracking System of Autonomous Vehicles

Joint Optimization of Model Partition and Resource Allocation for Split Federated Learning Over Vehicular Edge Networks

Federated Split Learning With Data and Label Privacy Preservation in Vehicular Networks

Blockchain for secure and efficient data sharing in vehicular edge computing and networks

Incentivizing Differentially Private Federated Learning: A Multi-Dimensional Contract Approach

Spears and shields: attacking and defending deep model co-inference in vehicular crowdsensing networks

AbstractVehicular CrowdSensing (VCS) network is one of the key scenarios for future 6G ubiquitous artificial intelligence. In a VCS network, vehicles are recruited for collecting urban data and performing deep model inference. Due to the limited computing power of vehicles, we deploy a device-edge co-inference paradigm to improve the inference efficiency in the VCS network. Specifically, the vehicular device and the edge server keep a part of the deep model separately, but work together to perform the inference through sharing intermediate results. Although vehicles keep the raw data locally, privacy issues still exist once attackers obtain the shared intermediate results and recover the raw data in some way. In this paper, we validate the possibility by conducting a systematic study on the privacy attack and defense in the co-inference of VCS network. The main contributions are threefold: (1) We take the road sign classification task as an example to demonstrate how an attacker reconstructs the raw data without any knowledge of deep models. (2) We propose a model-perturbation defense to defend against such attacks by injecting some random Laplace noise into the deep model. A theoretical analysis is given to show that the proposed defense mechanism achieves$$\epsilon$$ϵ-differential privacy. (3) We further propose a Stackelberg game-based incentive mechanism to attract the vehicles to participate in the co-inference by compensating their privacy loss in a satisfactory way. The simulation results show that our proposed defense mechanism can significantly reduce the effects of the attacks and the proposed incentive mechanism is very effective.

Hybrid sensor network with edge computing for AI applications of connected vehicles