Ford has been developing self-driving vehicles for about a decade, and many of the technologies used to create Ford's self-driving Fusion fleet are being added to new models as "driver assist" features. Take Pre-Collision Assist with Pedestrian Detection for example. Using sensors and cameras, just like a self-driving car, the 2017 Ford Fusion can identify the shape of a human being crossing the road, and automatically brake to avoid hitting them.

But fully self-driving cars may still be farther away than we think.

This month, researchers published a study on how malicious hackers could confuse autonomous vehicles into reading a Stop sign as a Speed Limit 45 sign—a potentially disastrous mistake—simply by altering the road signs.

The attackers would first need to know what algorithm the vehicle uses to identify different road signs, but once armed with that information, they can manipulate the signs using "adversarial perturbations" which are difficult to distinguish from run-of-the-mill graffiti by the human eye.

Fortunately, this type of research is exactly how engineers learn to make self-driving cars foolproof, and automakers are already working on solutions.

Categories: News, New Inventory