车载自组网中信任建模关键问题研究

Vehicular Ad-Hoc Network (VANET) contributes to the improvement of road safety and driving comfort. However, VANET is vulnerable to malicious behaviors and attacks, due to its unique characteristics. This project intends to conduct a comprehensive research on the key issues for trust modeling in VANET from a perspective of the full life cycle of trust, and propose corresponding trust models and solutions for initial trust establishment, trust evaluation and trust evolution. In concrete terms, this project will build an initial trust establishment model for neighboring nodes by aggregating multiple trust factors, and adequately address the challenge of trust initialization in VANET. Meanwhile, this project will present self-certified trust evaluation models under the Road Side Unit (RSU)-based and self-organized architectures, respectively, and provide effective methods for the mutual trust evaluation among nodes and recognition of malicious nodes in VANET. In addition, this project will propose a novel trust cascading theory and a lightweight trust cascading model, and provide strong support for distinguishing real and unreal messages in VANET. Furthermore, this project will put forward a reinforcement learning-based scheme in which the trust parameters can be optimized automatically, and provide practical methods for the automatic adjustment and optimization of trust models, and enable them to automatically adapt to the dynamic VANET environment. The research achievements will provide important theoretical basis and technical support for establishing the complete trust architecture in VANET, and effectively improve the security and reliability of VANET.

车载自组网有助于提高道路安全性和驾驶舒适性。然而由于其独特特性，车载自组网对于恶意行为和攻击是脆弱的。本项目拟从信任全生命周期的视角综合研究车载自组网中信任建模的关键问题，为初始信任建立、信任评估、信任演进分别提出相应的信任模型和解决方案：构建聚合多信任因素的邻近节点间初始信任建立模型，充分解决车载自组网中的信任初始化难题；分别提出在基于路侧单元和自组织两种体系架构下的自证明信任评估模型，为车载自组网中节点间的相互信任评估和恶意节点识别提供行之有效的方法；提出崭新的信任级联理论，并构建轻量级的信任级联模型，为车载自组网中的真假消息甄别提供强有力支撑；提出基于强化学习的信任参数自动优化方案，为信任模型的自动调整与优化提供切实可行的方法，并使其能够自动适应动态变化的车载自组网环境。研究成果将为构建车载自组网中完整的信任体系提供重要的理论依据和技术支撑，并有效提高车载自组网的安全性和可靠性。

车载自组网有助于提高道路安全性和驾驶舒适性。然而由于其独特特性，车载自组网对于恶意行为和攻击是脆弱的。本项目从信任全生命周期（即“信任初始化-->信任评估-->信任更新”）的视角综合研究车载自组网中信任建模的关键问题，为信任初始化、信任评估、信任更新分别提出相应的信任模型和解决方案：构建聚合多信任因素的邻近节点间信任初始化模型，充分解决车载自组网中的信任初始化难题；分别提出集中式架构下和分布式架构下的自证明信任评估模型，为车载自组网中节点间的相互信任评估和恶意节点识别提供行之有效的方法；提出崭新的信任级联理论，并构建轻量级的信任级联模型，为车载自组网中的真假消息甄别提供强有力支撑；提出基于强化学习的信任参数自动优化方案，为信任模型的自动调整与优化提供切实可行的方法，并使其能够自动适应动态变化的车载自组网环境。研究成果能够为构建车载自组网中完整的信任体系提供重要的理论依据和技术支撑，并有效提高车载自组网的安全性和可靠性。