革命性汽车行业发展 - 汽车自动驾驶

Article By : Arrow Electronics

The automatic driving of automobiles has been a new trend in the development of the automobile industry. Major automobile factories and information industry manufacturers are actively investing in the development of automatic driving vehicles, and many automatic driving cars have been actually driving on the roads. This article will introduce you to the development overview of automatic driving technology and related solutions.

安全是自动驾驶系统的绝对要求

自动驾驶汽车在许多科幻电影中出现了很长时间。随着科学和技术的快速发展，自动驾驶功能不再是梦想，而是现实中的。我们可以想象一个全自动的运输系统，车辆可以感知道路上的各种州，相互通信，并且也可以与交通信号通信。结果，交通事故的可能性将最小化（不可能实现零事故，因为“行人”也无法与自动化系统通信）。

从各种统计数据中可以发现，全球约有125万人每年死于交通事故，相当于每天3500人。这不包括因汽车事故而受到轻微或严重伤害的人数。在交通事故的原因中，大多数是由人类的疏忽，道路状况和环境因素引起的，因为导致交通事故的机械故障比例不高。这也意味着，如果将自动驾驶技术引入车辆，则可以有效减少事故的数量。尽管目前受到法律规范的限制，以及人类在自动驾驶中缺乏成熟和信任，但自动驾驶技术的发展仍然存在许多障碍。但是，随着自动驾驶技术的逐渐成熟度以及对道路驾驶的更实际验证，人们认为这将逐渐改变人们对自动驾驶技术的信心。

自动驾驶的原则是使用雷达,optical radar, GPS and computer vision and other technology to sense the environmental conditions, then through the advanced control system, the sensory data is converted into appropriate navigation roads, as well as obstacles and related signs, control the movement and action of the vehicle along the road to automatically get the vehicle to the set destination.

The key factor for people to accept and trust the autopilot system is “safety”. The safety of the autopilot is not an option nor a feature, but always an absolute pursuit. NVIDIA, which is actively involved in the development of the automatic driving processor, has launched the world's first single chip automatic driving vehicle processor DRIVE Xavier recently. From the very beginning, NVIDIA has incorporated safety into the design of NVIDIA DRIVE computers for automatic driving vehicles. Including hardware and software stack, experts have adopted security technology in all aspects of the computing system. All R&D tools and methods used by NVIDIA should ensure that the developed software achieves expected performance, is secure, reliable and backup; and the strict process is determined to ensure that there is no safety dead angle.

NVIDIA has invited TÜV SÜD, the world’s leading automotive safety and reliability company, to evaluate the safety concept of the newNvidia Xaviersystem-on-chip (SoC). “Safety first” computer design is equivalent to the combination of component expertise, architecture, design, tools, methods and best practices. TÜV SÜD evaluated the safety, durability and quality of cars in accordance with national and international standards and confirms that the Xavier system-on-chip architecture is suitable for automatically driven vehicle applications, it shows that Xavier has reached the functional safety standard ISO 26262 of the automotive industry and also achieved the highest functional safety rating in the automotive industry: ASIL-D.

Xavier是世界上第一个自动驾驶处理器，也是迄今为止最复杂的系统级芯片，有超过90亿个晶体管能够处理大量数据。Xavier的GMSL（千兆多媒体串行链路）高速IO将其连接到迄今为止最大的LIDAR，RADAR和相机传感器。系统级芯片中有六个处理器：ISP（图像信号处理器），VPU（视频处理单元），PVA（可编程视觉加速器），DLA（深度学习加速器），CUDA GPU和CPU。它每秒可以进行近400万亿个计算，仅人工智能的学习能力高达30万亿次。如今，它可以应用于最先进的生产工具。

为了协助汽车开发制造商开发自动驾驶系统，NVIDIA还推出了NVIDIA Jetson AGX Xavier开发人员套件。在Nvidia Jetson Agx Xavier开发人员套件的帮助下，很容易创建和部署自动驾驶系统的AI应用程序。

在Nvidia Jetpack，Deepstream SDK，Cuda®，Cudnn和Tensorrt软件库的支持下，此套件提供了您立即入门所需的所有工具。而且，由于它使用了新的Nvidia Xavier处理器，因此其性能比其前身产品Nvidia Jetson TX2高20倍以上，并且其能源效率高10倍。

Do you want to be absent from the application of the autopilot system? Please get to know NVIDIA’s autopilot vehicle processor DRIVE Xavier and Jetson AGX Xavier developer kit immediately, to accelerate the ideal world of auto driving together.