Unveiled 'Xavier,' a Drive Platform to Achieve SW-Based Autonomous Vehicles

NVIDIA showcased the autonomous driving capabilities of its NVIDIA DRIVE platform at GTC Japan, held for two days starting on the 12th, and stated that artificial intelligence will emerge as a key element of future transportation.

To prepare for changes in future transportation that will be driven by data-driven software technology, the industry is applying artificial intelligence computers to cars, trucks, and shuttle buses and conducting full-scale training using deep learning in data centers. These software-based systems have the advantage of being able to manage various levels of autonomous driving, ranging from Level II (partial automation) to Level V (full automation).

At GTC Japan, the final event of NVIDIA's GPU Technology Conference World Tour held in seven cities around the world this year, NVIDIA founder and CEO Jensen Huang explained how to implement the scalable architecture required for autonomous driving based on the NVIDIA Drive platform.

CEO Jensen Huang stated, “The future of transportation depends on software-based autonomous vehicles,” adding, “They will be equipped with a highly secure operating system based on sophisticated algorithms, and various types of applications will run on top of it.” Explaining that as the requirements for autonomous vehicles increase, automakers need scalable systems they can utilize, NVIDIA's R&D team showcased the training of autonomous test vehicles.

The next-generation NVIDIA Drive platform is based on Xavier, the world's most advanced system. CEO Jensen Huang stated that he hopes NVIDIA's Xavier will be supplied quickly to automakers and robot developers worldwide.

Autonomous vehicles equipped with Xavier can process external and internal sensor data using advanced neural networks, enabling eye tracking, gesture recognition, and natural language understanding, which allows for the implementation of applications based on 360-degree surrounding environment recognition.

On highways, Drive Zavier acts as an auxiliary driver, taking over the driving duties while fully recognizing the surrounding environment to execute functions such as cruise control, lane keeping, and automatic lane changes. The system's functionality can also be extended throughout the vehicle's lifespan through wireless software updates using Wi-Fi or cellular connections.

In addition, the NVIDIA Drive IX software development kit enables a smart experience in autonomous vehicles by allowing the car to notify the driver of potential hazards in the external environment and detect situations such as drowsy driving or decreased attention, thereby providing appropriate warnings.

CEO Jensen Huang explained, “Artificial intelligence will not only change the experience outside the vehicle, but will also completely transform how drivers enjoy driving and interact with the vehicle.”

Pioneer, a global manufacturer of automotive electronics, announced that it is collaborating with NVIDIA to apply the Drive platform to its 3D LiDAR sensors. Pioneer recently began supplying 3D LiDAR demonstration products to automotive manufacturers and ICT companies in Japan and around the world.

Tier IV, a Japanese developer of autonomous driving technology, unveiled 'Milee,' an electric golf cart capable of fully autonomous driving on the final mile using the NVIDIA Drive platform. Milee is equipped with all the necessary features for autonomous driving, including a 3D mapping positioning system using laser scanners, object detection, decision-making algorithms, and driving path mapping, and is designed for short-distance driving with a top speed of 20 km/h.

NVIDIA unveiled 'BB8,' a research autonomous vehicle containing the fundamental elements of autonomous driving, for the first time at GTC Japan, providing event participants with an opportunity to see firsthand the various sensors that enable autonomous driving and to reconsider the future of transportation brought about by artificial intelligence.