Go homepage(回首页) Upload pictures (上传图片) Write articles (发文字帖)
The author:(作者)aaapublished in(发表于) 2013/12/8 8:35:22 NVIDIA： super computers of cars on the road，
NVIDIA: super computers of cars on the road-NVIDIA, smart cars, sensor information-IT NVIDIA: super computer cars on the road
In an era of feature, Internet phone comes with features, while in the Smartphone era, could call the phone into the computer. Same change occurs in the car business?
High-performance system-level processors were desktop leaped on asphalt
The NVIDIA is headquartered in Santa Clara, California to create the graphics processor (GPU) is famous for. Its chips to improve the game-PC, laptop , workstation, and the efficiency of supercomputers. Which may come as a surprise is that it is pursuing cars become advanced networking mobile sensor-driven machines.
There are now 4 million cars strong built-in NVIDIA Tegra chip. The company also works with BMW, Rolls-Royce, Volkswagen/Audi, Aston Martin, and other famous brands have formed a partnership, thus there will be an additional 25 million vehicles using this chip. NVIDIA, with Japan and the United States automobile manufacturer cooperation can be expected soon.
NVIDIA automotive development kit is a box the size of a car stereo, provide Hyundai with all I/O connectors including USB, HDMI and Ethernet. NVIDIA system-level processors – is essentially a fully functional stand-alone computers---driven dashboard, navigation and infotainment systems. For electric car Tesla Model s dual touch-screen console provides computing power.
NVIDIA for gaming and technology designed to provide some of the most complex 3D rendering--so it's brand new car computing overwhelmingly of first-mover advantage. According to Gartner analyst at automotive information and communication technology diluo·kesiluofusiji (Thilo Koslowski), other companies--high-throughput, even Apple-will do a similar transition from consumer electronics chips to the automotive chips.
"But NVIDIA in pushing the envelope. "P.Koslowski said," it's like a battle between Ferrari and masses. ”
When the sensor meets the road
Though vivid digital dashboard is cool, but compared to the NVIDIA Tegra concept car level processing systems, they also only pediatrician. High performance processor allowing car companies to design a virtual car model on the screen, and then runs in a virtual wind tunnel precise or accurate aerodynamic simulation of diversionary force control system simulation of road test. Today, the constant expansion of technology is being used to interpret and integrate sensor data streams.
Current important vehicle components including cameras, radar, sonar, laser sensors, and so on.
"The CPU (central processing unit), GPU, graphics processor, audio processors and video processors all incorporated in the little things. "The NVIDIA advanced complexity andquality (Danny Shapiro), referring to embed the thumb nail size is slightly larger than the card circuit board components. The circuit board with the equipment needed to run like a stand-alone computer memory and components.
How to write a car
"It will run on Linux, Windows or Android system. "Shapiro explained," now, the software is based on this incredible processing power, and is able to offer car makers want any function. ”
NVIDIA offers a wide range of execution algorithm software library, thus helping car programmer and designers to accelerate development. NVIDIA provides software and hardware does not provide application--similar to the mining company's Entertainment Arts, Ubisoft, Valve game company.
NVIDIA UI Composer to speed up writing systems with built-in 3D objects such as gauges, and dials of the dashboard design. They sort of run-time engines.
Gaming graphics chip different is designed for entertainment, car rendering, and data interpreting system must be hundred percent accurate, because this is a matter of life or death.
For owners, the technology also means a higher degree of customization. If you don't like the 2014 Audi 3 dashboard designs? No problem. You can download from 1970 Audi 100 meter display device. The system directly from the car's CANBUS network for engine speed and performance information.
In addition, the big processor of cars means that car companies--is generally considered backward technology, will take three or four years to develop a new engine or transmission--will be able to begin to catch up with the pace of innovation in consumer electronics.
"Auto eject the old module, replace them with new ones. Per vehicle per year do not need to redesign can get a more powerful system. "Shapiro said," just like a phone. ”
Smart car
Richer we can explore this for in-car navigation systems and streaming capabilities means--cool stuff, but the core functionality of the car even worth thinking about, such as computer-based acceleration, braking and steering.
"The Holy Grail-class" technology is object detection and natural language processing. In 2013, the Las Vegas International Consumer Electronics Show (CES), Audi released a traffic assistance system, the system uses cameras and radar to detect traffic congestion, after securing permission from the owner of the shell starts advanced constant speed control function, maintain a safe distance from the car in front, while automatically turned to keep the drive in the lane. Heralding the NVIDIA vehicle platform development direction in the future.
On the session of the CES, NVIDIA, the Audi A7 navigation system shown on street view (Street View) feature integration. The company is expected to be returned to CES in January 2014, announced that the latest progress, car automation, which one step closer to reality.
The processing power required to imagine such a scenario coming: auto run dashboard processing unit according to your wishes, rendering speed, and engine features, hyper-realistic integration of Google Street view navigation system voice instructions guide you to your destination, and the kids are happy to watch Netflix online movies. Looking to the side of the camera to detect speed limit signs, and appear on the dashboard. Radar and LiDAR operation 30 to 60 times a second, to keep track of all traffic, the distinction between other cars with the rush in the middle of the road guy--let the car ready to brake if necessary.
Meanwhile, the in-car camera monitors the driver's status.
"Those with a similar process for pedestrian detection capability will able to blink detection, run the algorithm to determine whether drivers distracted or dozed off. "Shapiro notes that" who had got on the bus? Let us give that individual to adjust seats, radio and mirrors. "We're talking about smart cars.
This scenario is expected to soon become a reality, but the premise is to have a very fast and efficient processor.
"Our cars to develop a mobile supercomputers that can handle all these tasks. , "Shapiro said.
英伟达：把汽车做成马路上的超级电脑 - 英伟达,智能汽车,传感器 - IT资讯英伟达：把汽车做成马路上的超级电脑
在功能机的时代，上网只是手机的附带功能，而在智能手机时代，手机变成了能打电话的电脑。同样的变革会发生在汽车行业吗？
高性能系统级处理器正从桌面跃向柏油马路
总部位于加州圣克拉拉的英伟达以打造图形处理器（GPU）而著称。它的芯片提高了游戏型PC、笔记本电脑、工作站和超级计算机的效率。而可能让你感到意外的是，它也在推进汽车成为传感器驱动的先进联网移动机器。
目前已有400万辆汽车内置英伟达强大的Tegra芯片。该公司还与宝马、劳斯莱斯、大众/奥迪、阿斯顿马丁等知名汽车品牌达成了合作，因而还将有另外2500万辆汽车采用该芯片。英伟达与日本和美国的汽车厂商的合作也指日可期。
英伟达的汽车开发工具包是一汽车音响大小的盒子，提供现代汽车所需要的所有I/O连接器，其中包括USB、以太网和HDMI。英伟达系统级处理器——实质上是功能完备的独立计算机——驱动仪表板、导航系统和信息娱乐系统。也正是它为特斯拉的电动汽车Model S的双触摸屏控制台提供计算能力。
英伟达为游戏和技术设计提供部分最复杂的3D渲染技术——因此它在全新的汽车计算领域占据绝对的先发优势。据Gartner汽车信息与通信技术分析师蒂洛•科斯洛夫斯基（Thilo Koslowski）称，其它的公司——高通甚至是苹果——将进行类似的从消费电子芯片到汽车芯片的转型。
“不过英伟达在挑战极限。”科斯洛夫斯基说道，“这好比是法拉利与大众之间的较量。”
当传感器遇上道路
那些生动的数字仪表板纵然很酷，但相比英伟达构想的汽车级Tegra处理系统，它们还只是小儿科。高性能处理器可让汽车公司在屏幕上设计虚拟汽车模型，然后在虚拟风洞中运行精确的气体动力模拟或者进行准确的牵制力控制系统模拟路测。如今，该技术正被用于解读和整合传感器不断扩展的数据流。
当前的汽车的重要部件包括摄像头、雷达、声波定位仪、激光传感器等等。
“CPU（中央处理器）、GPU、图像处理器、音频处理器和视频处理器悉数并入这个细小的东西里。”英伟达汽车高级总监丹尼•夏皮罗（Danny Shapiro）谈到一嵌入略大于纸牌的线路板的拇指指甲大小组件。该线路板带有设备像独立计算机那样运行所需的内存和零部件。
如何编写汽车
“它将运行于Linux、Windows或者Android系统。”夏皮罗解释道，“现在，软件是基于这种难以置信的处理能力，因而能够提供汽车厂商想要的任何功能。”
英伟达还提供范围广泛的执行算法的软件库，因而能够帮助汽车编程员和设计师加速开发。英伟达提供软件和硬件，不提供应用——类似于该公司与Entertainment Arts、Ubisoft、Valve等游戏公司的关系。
英伟达的编写系统UI Composer可加速含计量表、转盘等内置3D物体的仪表板的设计。它们像是运行时引擎。
与为娱乐设计的游戏图形芯片不同的是，车载系统的渲染和数据解读都必须百分百精确，因为这是生死攸关的事情。
该技术对于车主来说还意味着更高程度的定制化。你如果不喜欢2014款Audi 3的仪表板设计？没问题。你可以去下载来自1970款Audi 100的仪表显示装置。该系统直接从汽车的CANBUS计算机网络获取引擎速度和性能信息。
此外，汽车的超级处理器意味着汽车公司——通常被认为技术落后，要三四年才能开发出新引擎或者传动装置——将能够开始赶上消费电子领域的创新步伐。
“给汽车弹出旧模块，换上新的。每辆汽车每年无需重新设计都可以获得更加强大的系统。”夏皮罗说道，“就像手机那样。”
智能汽车
我们可以探讨下这对于车载更加丰富的导航系统和流媒体功能意味着什么——酷玩意，但汽车的核心功能更加值得思考，如基于计算机的加速、刹车和转向。
“圣杯级”的技术是物体检测和自然语言处理。在2013年的拉斯维加斯国际消费电子展（CES）上，奥迪发布了交通阻塞援助系统，该系统利用摄像头和雷达检测交通拥挤情况，获得车主的批准后壳启动先进的恒速操纵功能，与前面的车保持一段安全距离，同时自动转向保持在车道线内行驶。这预示着英伟达车载平台的未来发展方向。
在那届CES上，英伟达在奥迪A7的导航系统中展示了对谷歌街景（Street View）功能的整合。该公司预计会在2014年1月回到CES公布最新的进展，使得汽车自动化进一步接近现实。
想象一下这种情景所需的处理能力即将到来：汽车的处理单元按照你的意愿运行仪表板、渲染速度和引擎功能，超逼真的整合谷歌街景的导航系统执行语音指令引导你到目的地，而孩子们则在开心地看Netflix的在线电影。瞄向路边的摄像头能够检测限速标志，并在仪表板上进行显示。雷达和激光雷达一秒钟运算30到60次，以追踪所有的交通状况，区分其它汽车跟冲在马路中间的孩子——让汽车准备好在必要时刹车。
与此同时，车内的摄像头负责监测驾驶者的状态。
“那些与用于行人探测的类似处理能力将能够进行眨眼侦测，运行算法确定驾驶者是否分心或者打瞌睡。”夏皮罗指出，“谁刚上车了？让我们给那个人调整座椅、广播和后视镜。”我们在说的是智能汽车。
这种情景有望不久后就成为现实，但前提是要有非常快速高效的处理器。
“我们在为汽车开发一个可处理所有这些任务的移动超级计算机。”夏皮罗说。
赞