NVIDIA® DriveWorks SDK
Advances in automotive technology continue to drive more computing power, more sensors, more functionality, more drivers assistance, and more autonomy into cars.
The NVIDIA® DriveWorks SDK provides everything software engineers need to develop Autonomous Vehicles (AV), such as reference applications, tools, and a library of modules for sensor data acquisition, point cloud processing, computer vision, deep learning, and more.
DriveWorks is optimized to achieve the full throughput limits of the computer and provides a run-time pipeline framework that spans sensing, perception, localization, planning, and actuation, enabling real-time end-to-end self-driving applications.
DriveWorks is the path to NVIDIA’s state-of-the-art AV research and development. Start developing advanced applications with DriveNet, LaneNet, and OpenRoadNet, NVIDIA’s proprietary DRIVE AV Perception Deep Neural Networks (DNNs). New and improved modules, neural networks, and functionality are delivered with each SDK release.
DriveWorks is designed to be modular, optimized, and flexible. You can use it out of the box, or enhance it with your own code to create your own AV solution.
DriveWorks is available as part of the NVIDIA DRIVE Software provided to select Automakers, Tier 1 Suppliers, and Research Institutions working on developing systems that enable cars to drive themselves.
A Few DriveWorks Features
DriveWorks Sensor Abstraction Layer
A unified interface to AV sensors that provides easy data capture and lifecycle management of camera, radar, lidar, IMU, GPS, and CAN sensors. The SAL also provides hardware engine support for efficient processing.
DriveWorks Algorithm Modules
Hardware-accelerated sensor data processing algorithms that include accumulation and registration (point clouds) and 2D tracking, rectification, and structure from motion (computer vision), and more!
DRIVE AV Perception DNNs
Proprietary deep neural networks generated with NVIDIA TensorRT provide advanced situational understanding. Detect objects such as cars, pedestrians, and bicycles using DriveNet. Detect lane markings with LaneNet. Detect drivable collision-free space using OpenRoadNet.
DriveWorks Tools & Samples
Tools that includes data capture (record, synchronize, and playback) and sensor calibration to assist developers in creating, analyzing, and maintaining Autonomous Vehicle software. Samples serve as a starting point for developing, porting, and optimizing AV applications.