Autonomous cars baffled by humans

September 4, 2015 | By | In News

Google’s self-driving cars are developed in such a complex way that they have now become baffled by human engagement, particularly when they come face-to-face with cyclists.

This is down to the vehicle’s default to caution. Therefore, when faced with human engagement in nonstandard behavior, strange incidents occur.

One such incident was reported earlier this month in Austin, Texas, when the robot car became extremely confused by a man riding a fixed-gear bike.

The Google car and the cyclist both arrived at a crossroads. The car got there a fraction of a second before the bike, and the cyclist says he waited for it to continue through.

But, instead of just putting a foot on the ground at the stop sign, the cyclist did what's called a trackstand – a common manoeuvre among fixed-geared riders, which involves the rider pedaling both forward and backward while trying to stay upright and moving a minimal distance.

The issue is that, during such a trackstand, the rider does usually move slightly forward or backward – at least enough to alert the conscientious Google car that a human might be blasting through the stop sign.

The cyclist told Business Insider: “It apparently detected my presence and stayed stationary for several seconds. It finally began to proceed, but as it did, I rolled forward an inch while still standing. The car immediately stopped.”

It was this motion that confused the car, as it could not understand what the cyclist was doing.

He added: “We repeated this little dance for about two full minutes.”

But even though the car’s behaviour was strange, the cyclist says he felt safer dealing with an autonomous vehicle than a human-operated car.

Google have patented a method whereby self-driving cars can identify cyclists and understand their hand signals

Photo: Martialred

Social media & sharing icons powered by UltimatelySocial

By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close