摘要
传统车机系统在算力扩展与生态集成方面存在结构性瓶颈。文章面向弱终端强云端的架构模式构建了以编解码显示为车端基础、算力与应用完全云托管的云车机方案,设计了云端GPU弹性调度、边缘计算部署、指令优先级策略与低延时视频传输机制,并构建车端与云端之间的实时双向控制通道。结果表明该架构具备显著的算力解耦能力与生态适配灵活性,对智能座舱向平台化、低成本演进具有支撑意义。
There are structural bottlenecks in the expansion of computing power and ecological integration of traditional vehicle-machine systems.This paper constructs a cloud vehicle machine solution with codec display as the foundation of the vehicle,computing power and application fully managed in the cloud,designs the cloud GPU elastic scheduling,edge computing deployment,instruction priority strategy and low-latency video transmission mechanism for the architecture mode of weak terminals and strong clouds,and constructs a real-time two-way control channel between the vehicle and the cloud.The results show that the architecture has significant computing power decoupling ability and ecological adaptation flexibility,which is of supporting significance for the evolution of intelligent cockpits to platformization and low cost.
作者
权香妮
蔡勇
韦建平
Quan Xiangni;Cai Yong;Wei Jianping
出处
《时代汽车》
2025年第24期117-119,共3页
Auto Time
关键词
云车机
车云协同
音视频串流
交互控制
Cloud Car Machine
Vehicle-cloud Collaboration
Audio and Video Streaming
Interactive Control
