• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 人工智能与数据挖掘 • 上一篇    下一篇

适应于自动驾驶的计算结构与平台综述

邹文超,李仁发,吴武飞   

  1. (湖南大学信息科学与工程学院,湖南 长沙 410082)
  • 收稿日期:2018-01-12 修回日期:2018-07-11 出版日期:2019-03-25 发布日期:2019-03-25
  • 基金资助:

    国家重点研究和发展计划(2016YFB0200405,2012AA01A30101);国家自然科学基金(61672217)

Summary on computing structure and
platform for autonomous driving

ZOU Wenchao,LI Renfa,WU Wufei   

  1. (College of Computer Science and Electronic Engineering,Hunan University,Changsha 410082,China)
  • Received:2018-01-12 Revised:2018-07-11 Online:2019-03-25 Published:2019-03-25

摘要:

信息与通信技术领域的不断发展,推动了自动驾驶汽车的巨大进步。自动驾驶汽车有效解决了用户对出行日益提高的安全、便捷、舒适、高效等个性化需求,有着巨大的商业前景和应用价值。首先从自动驾驶的定义出发,对自动驾驶汽车计算任务进行了分类总结,阐述了自动驾驶汽车集移动端、边缘节点和云端一体的计算结构。然后分类讨论了目前自动驾驶汽车移动端计算平台,呈现出异构多核的特点,并对当前主流的底层计算平台进行了对比分析。最后,从性能与成本和信息安全两个方面总结目前自动驾驶汽车计算结构和平台遇到的问题与挑战。
 

关键词: 自动驾驶, 无人驾驶, 人工智能, 高性能计算

Abstract:

With the continuous development of information and communication technology, the field of autonomous driving cars has made a great progress. Autonomous driving cars effectively solve the user’s increasing demand for safety, convenience, comfort, efficient travel, and other personalized needs. The autonomous driving car is very popular among the traditional manufacturers, emerging auto companies and people for its great commercial prospect and application value. Starting from the definition of autonomous driving cars, we classify the computing tasks of autonomous driving cars and describe the computing structures of the mobile terminal, the edge node and the cloud. Then, we discuss current mobile terminal computing platforms of autonomous driving cars, which shows heterogeneous and multi-core characteristics, and  compare and analyze the current mainstream underlying computing platforms. Finally, we summarize the problems and challenges encountered in the computing structure and platform of autonomous driving cars from aspects of performance cost and information security.
 
 

Key words: autonomous driving, unmanned driving, artificial intelligence, high performance computing