期刊文献+

智能驾驶-AI芯片的算力研究 被引量:2

Intelligent Driving–AI Chips Computing Power Research
下载PDF
导出
摘要 在一汽对红旗汽车智能化工作推进的背景下,L3级有条件自动驾驶及L4、L5级全自动驾驶提上工作日程。自动驾驶功能的实现离不开具有强大计算能力的硬件平台,而硬件计算平台的核心则是负责完成海量传感器数据采集、处理、分析工作的AI性能芯片。本文对AI芯片的计算能力表示单位及算力计算公式进行解释说明,同时对自动驾驶硬件方案所采用的主流AI芯片的算力进行分析统计,最后对AI芯片的应用选型提出建议。 Under the background of promotion program for FAW's connected and intelligent HongQi limousine, the L3 of conditional automation,L4 and L5 automatic driving are put on the development schedules. The realization of the autopilot function is inseparable from the hardware platform with powerful computing power. However,The key component of the hardware computing platform is the AI performance chip, which is responsible for the massive sensor data acquisition, processing and analysis. This paper explains the unit of calculating processing capacity and the formula of calculating for the AI chip, analyzes and counts the processing capacity of the AI chip used in the autopilot hardware scheme. In the end authors put forward some suggestions for application and selection of AI chips.
作者 于继成 王强 赵目龙 焦育成 Yu Jicheng;Wang Qiang;Zhao Mulong;Jiao Yucheng(Intelligent and Connected Vehicle Development, China FAW Group Corporation Limited, Changchun 130011)
出处 《汽车文摘》 2019年第1期31-36,共6页 Automotive Digest
关键词 智能驾驶 算力单位 算力计算方法 算力对比分析 Intelligent Driving Computing Capacity Unit Computing Method Computing comparative Analysis
  • 相关文献

同被引文献17

引证文献2

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部