Sign In

Communications of the ACM

ACM TechNews

Programming Safety Into Self-Driving Cars

View as: Print Mobile App Share:
A vision for urban transportation in 2030: the GM-Segway EN-V Laugh (Xiao) exhibited at the Shanghai Expo 2010.

University of Massachusetts, Amherst professor Shlomo Zilberstein is investigating ways to help semi-autonomous systems make better decisions.

Credit: Wikimedia Commons

With the help of funds from the U.S. National Science Foundation, University of Massachusetts, Amherst professor Shlomo Zilberstein has been investigating ways of helping semi-autonomous systems make better decisions.

A key focus of Zilberstein's recent work has been autonomous vehicles. Zilberstein, like many in the field, believes an era of fully autonomous vehicles that will operate more or less independently of human passengers, is likely far off. Instead, he sees a long period were humans act as co-pilots or supervisors to autonomous vehicles, and his research focuses on how to make that relationship more functional.

Autonomous vehicles have repeatedly been shown to have difficulty making timely decisions in ambiguous situations such as four-way stops where picking up on social cues is important. One of Zilberstein's new algorithms seeks to ameliorate this issue by allowing semi-autonomous systems to make sequential decisions in situations where there are multiple objectives. Zilberstein provides an example of a scenario wherein a driver is fatigued and a vehicle using his algorithm will be able to autonomously plot a route home based on pre-existing priorities such as speed and directness, or ease of driving.

From National Science Foundation
View Full Article


Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account