车联网场景下数字孪生网络架构及关键技术研究

Research on Digital Twin Network Architecture and Key Technologies in Vehicular Networking Scenarios

智能化已成为未来车辆的核心特征之一,通过在网络端建立数字孪生体,能够满足车辆智能化的多重需求。一方面,数字孪生体为车辆提供了非实时智能化辅助管理,另一方面,它支持车辆实时智能化辅助驾驶。该研究针对这一需求,对智能车辆的两大类智能场景进行了深入分析,并提出了一种车联网场景下数字孪生网络新型架构,还对车辆数字孪生体的感知构建、管理、迁移、联邦学习、仿真优化决策等关键技术进行了详尽分析,以解决孪生体孪生感知、高效管理、随车迁移和模型通用泛化等方面的挑战,为运营商服务智能车辆提供了有效、可行的实现方法和路径。

Intelligence has become one of the core features of future vehicles,and by establishing a digital twin on the network side,it can meet the multiple needs of vehicle intelligence.On the one hand,the digital twin provides non-real-time intelligent assisted management for vchicles,and on the other hand,it supports real-time intelligent assisted driving for vehicles.The study addresses this ned,provides an in-depth analysis of two major types of intelligent scenarios of intelligent vehicles,and proposes a novel architecture of digital twin network in vehicle networking scenarios,and also thoroughly analyzes the key technologies of perception construction,management,migration,federated learning,and simulation-optimized decision making of digital twins for vehicles,in order to solve the twin perceptions,efficient management,migration with the vehicle,and generalized generalization of models of the twins and other challenges,providing effective and feasible realization methods and paths for operators to serve intelligent vehicles.