NVIDIA DriveWorks SDK

NVIDIA DriveWorks provides a Sensor Abstraction Layer that supports capturing of data from various sources. It is designed to provide the following features:

The Image Processing Library provides structures and algorithms to efficiently process image data. It features the following algorithms:

The DriveWorks Point Cloud Processing Library is built specifically for Point Cloud data as transmitted by LiDAR sensors. It provides the following GPU-accelerated algorithms:

The Calibration Module supports Dynamic Calibration for Camera, Radar, LiDAR and IMU sensors that are compatible with the DriveWorks Sensor Abstraction Layer.

Dynamic Calibration is a process where sensor parameters are re-estimated at runtime based on sensor measurements and vehicle motion. This process compensates for the effects of environmental changes or mechanical stress—such as changes in road gradient, tire pressure, or vehicle loading—which can affect the extrinsic parameters (position and orientation) of sensors during a vehicle's operation.

The Egomotion module uses a motion model to track and predict the vehicle’s pose based on sensor inputs.

It supports two types of motion models: an odometry-only model and, if an IMU is available, a model based on IMU and odometry. During run-time, the module takes measurements as input and internally updates the current estimation of the vehicle pose. The module can be queried for vehicle motion between any two points in time.