Abstract

This document describes techniques that enable a computing system to manage artificial intelligence (AI) and machine learning (ML) workloads in software defined vehicles (SDVs) through a unified, modular operating system for SDVs. SDVs may take advantage of distributed processing and hardware abstraction through virtualization, using standardized interfaces (e.g., virtio) to allow applications to run on any compliant virtual machine (VM) or Electronic Control Unit (ECU) within the vehicle. For instance, the vehicle’s internal architecture may include a multi-node (or multi-VM) AI workload orchestration framework that distributes tasks to available hardware, supports dynamic service deployment, and facilitates updates. The computing system may offload tasks between different systems, such as from an In-Vehicle Infotainment (IVI) system to a dedicated SDV node with specialized AI accelerators. Standardization of vehicle sensor signals for AI use cases may also help ensure model compatibility across different vehicle models. Further capabilities may include the integration of various AI technologies, such as imaging AI and specialized models running on edge devices. In this way, the technology may provide a standardized, scalable, and hardware-agnostic platform for deploying and managing AI workloads within SDVs.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS