Should driverless cars make life-or-death decisions?

An interesting article ran recently in Automotive News titled, “Should driverless cars make life-or-death decisions?” Author Keith Naughton asks some interesting questions and points out some of the difficult decisions engineers are now beginning to make. Read the short article at: http://www.autonews.com/article/20150625/OEM06/150629944?template=mobile .

This is not directly related to work zone ITS, but raises additional questions that most certainly are. When is it better for robotic cars to drive through a line of cones? Is it better to crash into a truck-mounted attenuator and risk injuring the operator? Or veer into a closed work zone where workers may be present?

The problem, as I understand it, is that the robots can’t go through a decision tree to decide what to do. As this video points out ( https://www.youtube.com/watch?v=IanxRZMIC-E ) there just isn’t enough time. Instead, programmers must build in the ability to react. The robot in the video hasn’t been taught when to grab the bar. Rather, it is taught how to react instead.

We will soon be enjoying significantly reduced fatality rates thanks to autonomous vehicles. But robots will now be making these decisions. In the past drivers involved in crashes were often tormented second guessing their split second decisions prior to the crash. Now we can make those decisions in advance in the way we program these systems. But they won’t always be easy decisions.