Introduction to DriveOS SDK APIs#

DriveOS APIs equip developers with a powerful toolkit to fully harness NVIDIA’s DriveOS platform for autonomous vehicles. For high-performance AI tasks, the CUDA API delivers GPU acceleration, supporting real-time applications for computer vision and data processing.The Deep Learning and Inference (DLI) APIs enable integration of deep neural networks (DNN) and large language models (LLMs), facilitating natural language processing and in-vehicle conversational AI. Core APIs like NvSIPL and NvStreams enable efficient data handling from various sensors, while the DriveWorks API provides essential tools for perception, mapping, and localization. Additionally, Safety and Security APIs ensure compliance with stringent automotive safety standards. These APIs collectively provide a scalable, high-performance foundation for developing advanced autonomous vehicle applications.

DriveOS SDK APIs#

DriveOS

API Description

Graphics

OpenGL, OpenGL-ES, Vulkan, Vulkan-SC

Windowing

X11, Wayland, EGLDevice/EGLOutput, Screen

NvDisplay

Audio

Camera

NvMedia

NvSIPL

NvStreams

EGLStreams

CUDA

TensorRT

LLM SDK

PVA SDK

Networking

DRIVE Update

NvLogging

Hypervisor Logging

Target Container Support

CAN Driver

DriveWorks

System Task Manager (STM)

cuDNN