Abstract: As the brain of intelligent connected vehicles, the computing platform mostly adopts multi-heterogeneous hardware architecture and uses multiple operating systems to run the environment perception, path planning and vehicle control algorithms with different levels of functional safety and cybersecurity. A virtualized software architecture, adapting to a variety of hardware platforms and multiple application scenarios is proposed in this paper. This architecture enables the resource sharing and partitioning isolation between virtual machines, and provides a variety of fault tolerance and protection mechanisms to concurrently meet the requirements of functional security and cybersecurity. In addition, an efficient communication mechanism between virtual machines and a virtualized performance optimization mechanism are also proposed and verified, which can meet the requirements of computing platform of intelligent connected vehicles

Key words: intelligent connected vehicles, computing platform, operating system, virtualization