Intelligence testing for autonomous vehicles: A new approach

Li Li; Wu Ling Huang; Yuehu Liu; Nan Ning Zheng; Fei Yue Wang

doi:10.1109/TIV.2016.2608003

Intelligence testing for autonomous vehicles: A new approach

Li Li
, Wu Ling Huang
, Yuehu Liu
, Nan Ning Zheng
, Fei Yue Wang

College of Artificial Intelligence and Robotics

Tsinghua University
Jiangsu Province Collaborative Innovation Center of Modern Urban Traffic Technologies
Chinese Academy of Sciences
Xi'an Jiaotong University

Research output: Contribution to journal › Article › peer-review

275 Scopus citations

Abstract

In this paper, we study how to test the intelligence of an autonomous vehicle. Comprehensive testing is crucial to both vehicle manufactories and customers. Existing testing approaches can be categorized into two kinds: scenario-based testing and functionality-based testing. We first discuss the shortcomings of these two kinds of approaches, and then propose a new testing framework to combine the benefits of them. Based on the new semantic diagram definition for the intelligence of autonomous vehicles, we explain how to design a task for autonomous vehicle testing and how to evaluate test results. Experiments show that this new approach provides a quantitative way to test the intelligence of an autonomous vehicle.

Original language	English
Pages (from-to)	158-164
Number of pages	7
Journal	IEEE Transactions on Intelligent Vehicles
Volume	1
Issue number	2
DOIs	https://doi.org/10.1109/TIV.2016.2608003
State	Published - Jun 2016

Keywords

Autonomous vehicles
Intelligence testing

Access to Document

10.1109/TIV.2016.2608003

Cite this

@article{e943c2a0094f4b919a64351cf0fc0215,

title = "Intelligence testing for autonomous vehicles: A new approach",

abstract = "In this paper, we study how to test the intelligence of an autonomous vehicle. Comprehensive testing is crucial to both vehicle manufactories and customers. Existing testing approaches can be categorized into two kinds: scenario-based testing and functionality-based testing. We first discuss the shortcomings of these two kinds of approaches, and then propose a new testing framework to combine the benefits of them. Based on the new semantic diagram definition for the intelligence of autonomous vehicles, we explain how to design a task for autonomous vehicle testing and how to evaluate test results. Experiments show that this new approach provides a quantitative way to test the intelligence of an autonomous vehicle.",

keywords = "Autonomous vehicles, Intelligence testing",

author = "Li Li and Huang, \{Wu Ling\} and Yuehu Liu and Zheng, \{Nan Ning\} and Wang, \{Fei Yue\}",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.",

year = "2016",

month = jun,

doi = "10.1109/TIV.2016.2608003",

language = "英语",

volume = "1",

pages = "158--164",

journal = "IEEE Transactions on Intelligent Vehicles",

issn = "2379-8858",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "2",

}

TY - JOUR

T1 - Intelligence testing for autonomous vehicles

T2 - A new approach

AU - Li, Li

AU - Huang, Wu Ling

AU - Liu, Yuehu

AU - Zheng, Nan Ning

AU - Wang, Fei Yue

PY - 2016/6

Y1 - 2016/6

N2 - In this paper, we study how to test the intelligence of an autonomous vehicle. Comprehensive testing is crucial to both vehicle manufactories and customers. Existing testing approaches can be categorized into two kinds: scenario-based testing and functionality-based testing. We first discuss the shortcomings of these two kinds of approaches, and then propose a new testing framework to combine the benefits of them. Based on the new semantic diagram definition for the intelligence of autonomous vehicles, we explain how to design a task for autonomous vehicle testing and how to evaluate test results. Experiments show that this new approach provides a quantitative way to test the intelligence of an autonomous vehicle.

AB - In this paper, we study how to test the intelligence of an autonomous vehicle. Comprehensive testing is crucial to both vehicle manufactories and customers. Existing testing approaches can be categorized into two kinds: scenario-based testing and functionality-based testing. We first discuss the shortcomings of these two kinds of approaches, and then propose a new testing framework to combine the benefits of them. Based on the new semantic diagram definition for the intelligence of autonomous vehicles, we explain how to design a task for autonomous vehicle testing and how to evaluate test results. Experiments show that this new approach provides a quantitative way to test the intelligence of an autonomous vehicle.

KW - Autonomous vehicles

KW - Intelligence testing

UR - https://www.scopus.com/pages/publications/85017288682

U2 - 10.1109/TIV.2016.2608003

DO - 10.1109/TIV.2016.2608003

M3 - 文章

AN - SCOPUS:85017288682

SN - 2379-8858

VL - 1

SP - 158

EP - 164

JO - IEEE Transactions on Intelligent Vehicles

JF - IEEE Transactions on Intelligent Vehicles

IS - 2

ER -

Intelligence testing for autonomous vehicles: A new approach

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this