Google Toes The Line Between Safe And Realistic With Autonomous Cars

Self-driving vehicles have been a closely watched development in the tech space as companies like Google Inc GOOG GOOGL and Tesla Motors Inc TSLA all work to come up with the safest and most advanced pilot version. While the cars aren't expected to hit showrooms in the near future, debates over whether or not such vehicles will be considered safe have already begun to bubble up among investors.

However, in an effort to plan for every eventuality, Google recently revealed that cars which are too safe may not be road-ready.

Related Link: Google Hires TrueCar President For Self-Driving Car Project

How Safe Is Too Safe?

Google's initial autonomous cars were programmed to drive in the safest way possible. That meant light breaking at any sign of danger and wide turns to avoid unseen pedestrians. While such measures may seem reasonable for a car that drives itself, Google says those measures actually made the car riskier. As most drivers don't behave in those ways, overly safe autonomous cars are creating more accident risks because regular drivers are unable to anticipate what the car is doing as such behavior isn't the norm on the road.

Humanizing The Cars

For that reason, Google says its next task is humanizing the way the cars drive.

While the car is able to avoid accidents through the use of sensors, developers must also take into account a self-driving car's potential to create accidents as well. Frequent, light braking may be a good way to keep an autonomous car from breaking the speed limit or avoiding danger, but other drivers may see the break lights and slam on their own breaks in response. Problems like these have led Google developers to study the habits of average drivers and work some of those traits into the cars' algorithms. That way, autonomous cars send other drivers the same signals a normal car would.