Sign In

Communications of the ACM

ACM TechNews

DARPA Assured Autonomy Seeks to Guarantee Safety of Learning-Enabled Autonomous Systems


DARPAs 2005 Grand Challenge.

The U.S. Defense Advanced Research Projects Agency has announced a new research program called Assured Autonomy, which aims to advance the ways computing systems can learn to better manage variations in the environment and enhance the predictability of autonomous systems like driverless vehicles, which are descendants of research pioneered more than a decade ago in the agency's 2005 Grand Challenge.

Credit: U.S. Defense Advanced Research Projects Agency

Researchers at the U.S. Defense Advanced Research Projects Agency have launched Assured Autonomy, a new research program that aims to advance the ways computing systems can learn and evolve to better manage variations in the environment and enhance the predictability of autonomous systems such as autonomous vehicles.

One new approach is based on the idea that autonomous systems must be at least as safe as comparable manned systems, known as "equivalent levels of safety." This approach compares known rates of safety incidents of manned systems and conducts physical trials to determine the corresponding incident rate for autonomous systems. However, studies indicate that assuring the safety of autonomous systems in this manner alone is prohibitive because it requires millions of physical trials that could take decades to complete.

The Assured Autonomy program's researchers will seek to develop tools that provide evidence a system can satisfy stated functional and safety goals.

From DARPA News
View Full Article

 

Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA


 

No entries found