The focus of self-driving technology development in 2016 is the development of sensing technology, so that self-driving vehicles can sense the surrounding environment and make independent judgments. In contrast, the focus of self-driving technology development in 2017 will focus on the construction of high-resolution maps (HD Maps), which are high-resolution maps with a resolution of up to the centimeters, allowing safe and efficient vehicles to travel safely and efficiently. It is crucial.

Su Jiaxing, NVIDIA's technical marketing manager, said that the first condition for self-driving is to equip the vehicle with the ability to perceive the surrounding environment, including the cars, people, and traffic signs. However, as self-driving vehicles are already aware of these objects, the next step is to supplement the high-resolution maps to further enhance the performance of the self-driving system.

In fact, the current navigation map not only has a slower positioning speed, but also has insufficient information for the self-driving system. If you want to judge whether the car should accelerate or decelerate or change lanes through a computer, the computer must know the car itself. There are several lanes and other information near the location and the car. Therefore, high-resolution maps may become an important development focus for self-driving cars this year.

Self-driving cars that already have the ability to sense and judge the surrounding environment can be used with high-resolution maps to know the location of each lane and accurately determine the correct position of the car itself. Thereby, the system can further determine whether it is necessary to pass or change lanes.

The process of collecting high-resolution images can be carried out through equipment such as Lidar, camera, etc., but the implementation process is not just to drive the car through the route, that is, the ability to display the map, the only way to collect the data first, and then A series of operations are required to create a 3D map.

The NVIDIA DRIVE PX2 is a scalable architecture that includes the AI ​​artificial intelligence required for the entire autonomous driving technology. At the GTC Technology Conference, NVIDIA released the DRIVE PX 2 AutoCruise, a technology designed for highway autopilots that continuously positions and maps. In addition, the company also launched the DriveWorks Alpha1 home-made operating system to support self-driving, providing a full range of auto-driving features, including detection, positioning, planning routes and motion response.

At present, NVIDIA has cooperated with several graphic manufacturers, such as Baidu, TomTom, Japan ZENRIN, and European HERE. Since the establishment of high-resolution maps is a very expensive investment in time and money, NVIDIA will continue to cooperate with other home map manufacturers in the future to jointly create high-resolution map information in different corners of the world, paving the way for the development of self-driving vehicles.

R&M BOX

Shenzhen Aierbaita Technology Co., Ltd. , https://www.aierbaitavape.com

Posted on