A critique of the SAE conditional driving automation definition, and analyses of options for improvement

• Published in: Cognition, technology & work Vol. 21; no. 4; pp. 569 - 578
• Main Authors: Inagaki, Toshiyuki, Sheridan, Thomas B.
• Format: Journal Article
• Language: English
• Published: London Springer London 01-11-2019; Springer Nature B.V
• Subjects: Aerospace Technology and Astronautics; Automation; Automobile driving; Automotive Engineering; Autonomous vehicles; Cognitive Psychology; Computer Science; Industrial and Organizational Psychology; Medicine/Public Health; Original Article; Traffic accidents & safety; User Interfaces and Human Computer Interaction; Request to intervene; Trading of authority; Automated driving; Levels of automation; Levels of driving automation; Shared control
• ISSN: 1435-5558; 1435-5566
• DOI: 10.1007/s10111-018-0471-5

The Society of Automotive Engineers defines five
levels of driving automation
(LoDA) (plus a “no-automation” level 0). Among them, the third level, called “conditional driving automation,” here denoted LoDA 3, performs the complete
dynamic driving task
(DDT) within a limited operational domain. Although the driver is free from any driving task while the automation is engaged, she is expected to be receptive to an automation-issued
request to intervene
(RTI) and is also expected to perform DDT fallback in a timely manner. This paper gives a method to derive an optimal design for RTI and proves that LoDA 3 coupled with the optimal RTI should never be simply called “conditional driving automation.” This means that the definition of LoDA 3 is not complete and that at least one important level is missing in the list for LoDAs. This paper provides two ways to resolve the problem.