Follow
Subscribe

NVIDIA Announces ‘DRIVE AP2X' 9.0, its Latest Level-2 Autonomous Driving Platform

Home > News > Content

【Summary】NVIDIA CEO Jensen Huang delivered the opening keynote at this week’s annual GPU Technology Conference in San Jose, California to announce some of the company’s latest products and breakthroughs in fields such as autonomous driving, artificial intelligence and deep-learning. Among the highlights this year was the announcement of Nvidia’s latest autonomous driving platform called DRIVE AP2X, a complete level-2 autonomous driving platform.

Eric Walz    Mar 21, 2019 6:44 PM PT
NVIDIA Announces ‘DRIVE AP2X' 9.0, its Latest Level-2 Autonomous Driving Platform
NVIDIA CEO Jensen Huang presenting at the GPU Technology Conference this week in San Jose, Calif.

NVIDIA CEO Jensen Huang delivered the opening keynote at this week's annual GPU Technology Conference in San Jose, California to announce some of the company's latest products and breakthroughs in fields such as autonomous driving, artificial intelligence and deep-learning.

Among the highlights this year was the announcement of Nvidia's latest autonomous driving platform called DRIVE AP2X, a complete level-2 autonomous driving platform. DRIVE AP2X includes Nvidia's DRIVE AutoPilot software, DRIVE AGX and DRIVE Works validation tools.

The AP2X platform is part of the open, flexible NVIDIA DRIVE platform, which is being used by hundreds of companies worldwide to build autonomous driving vehicles. The platform is slowly becoming essential hardware for global automakers working on autonomous driving, including Tesla, Audi, Mercedes Benz and Volvo.

The DRIVE AGX platform, which is included in the new AP2X release, is used to process data from the vehicle's cameras, lidar, radar, and ultrasonic sensors to understand the complete 360-degree environment in real-time. DRIVE AutoPilot integrates enhanced perception, visualization and mapping capabilities into one release, so a self-driving car can localize itself on a map and plan a safe path forward.

AP2X also adds a wide range of new autonomous driving capabilities, including more deep neural networks (DNNs), facial recognition capabilities and additional sensor integration options.

DRIVE AP2X incorporates the complete NVIDIA DRIVE AV software stack, which powers all of the necessary functions for fully autonomous driving. DRIVE AV includes perception, localization and path planning. Also included in AP2X is DRIVE IX intelligent cockpit experience, which includes an AI assistant.

Each software product runs on Nvidia's high-performance, energy-efficient NVIDIA Xavier system-on-a-chip (SoC). Xavier is Nvidia's first SoC designed specifically for autonomous driving. Nvidia spent four years and $2 billion in research to develop Xavier.

For developers, DriveWorks provides the reference applications, tools, and a comprehensive library of modules that take advantage of the robust computing power of the DRIVE AGX platform. AP2X also utilizes DriveWorks acceleration libraries and DRIVE OS, a real-time operating system (OS).

To enhance mapping and localization, DRIVE AP2X software will include MapNet, a DNN that identifies road lanes and landmarks. MapNet includes a suite of three distinct path planning DNNs to provide greater accuracy in keeping an autonomous vehicle precisely in a road lane.

dims (1).jpg

NVIDIA's Xavier system-on-a-chip

ClearSightNet is part of Nvidia's camera-based obstacle perception software, which allows the vehicle to detect camera blindness, for example when the sunlight shines directly into the sensor, or when rain or snow limits its vision. This allows the car to take action to make up for any sensor obstructions.

Clearsight net performs DNN inference on a live camera feed, evaluating each frame to detect camera blindness.

For driver monitoring, a new DNN enables facial recognition. Developers and manufacturers can use this feature to enable face identification as a passenger enters the vehicle. This can be used to unlock the doors, start the car, adjust the seats or set the climate control to a driver's preference when the DNN recognizes a driver's face.

DRIVE AP2X software will also feature new visualization capabilities that can be displayed to passengers. This is designed to help build trust for passengers riding in a self-driving car. It provides occupants a visualization of what the car's surround camera sees, along with the current speed, speed limit and driver monitoring, all displayed on one screen.

nvidis sff.jpg

NVIDIA's  "Safety Force Field"

NVIDIA also detailed the DRIVE Planning and Control software layer as part of its DRIVE AV software suite.

This layer consists of a route planner, a lane planner and a behavior planner that work together to enable a safe and comfortable driving experience. A primary component of the DRIVE Planning and Control software is NVIDIA Safety Force Field (SFF). It functions like a protective barrier surrounding the vehicle, keeping occupants safe.

SFF software analyzes and predicts the dynamics of the surrounding environment. It takes in sensor data and determines a set of actions to protect the vehicle and other road users.

A preview of DRIVE AP2X was on display this week at GTC.

DRIVE AP2X Software 9.0 will be released in Q2 2019.

Prev                  Next
Writer's other posts
Comments:
    Related Content