This content is not included in your SAE MOBILUS subscription, or you are not logged in.
A Tenable and Safe Approach to Autonomous System Development
ISSN: 0148-7191, e-ISSN: 2688-3627
Published April 15, 2021 by SAE International in United States
With over 36,000 deaths in the US per year the creation of autonomous vehicles is surely a worthwhile effort. To meet this challenge the autonomous vehicle industry has undertaken a very important and mind-boggling complex and difficult task. So difficult it may be the most complex engineering effort undertaken by humankind to this point. An effort that involves nascent machine and deep learning technology and approaches. It also largely relies on the use of public domain shadow and safety driving augmented by gaming architecture and approach centered simulation and modeling. These areas are the focus of this presentation. Areas that are untenable from a time, cost and safety standpoint. So untenable it will mean the demise of the industry, no public roadway L4/5 vehicles, no lives saved as a result and thousands of injured and fatally injured “safety drivers”. Finally, the paper will a discuss the resolution to these issues. A flipping of the current real-world to proper simulation paradigm. This paper will go in to details regarding the factors involved in the areas of concern and the recommended solution. These include time to regain proper situational awareness during handover, the need for many accident scenarios to be experienced, many times over, in order for them to be learned and technical and engineering simulation and modeling approach deficiencies with gaming architectures. And finally, a discussion on how the capabilities of DoD/aerospace simulation resolve these issues. (Note-this effort also relates to the air autonomous side.)