autoevolution
 

Carnegie Mellon Robotics Lab Director Says Self-Driving Cars Are Not Even Close

Autonomous cars appear to be around the corner, if we look at some of the players in the industry, but not all parties agree that driverless vehicles are ready to replace humans.
Modified Volvo XC90 for Uber 11 photos
Photo: Uber
Volvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeVolvo XC90 Drive Me test vehicle autonomous driving prototypeModified Volvo XC90 for UberModified Volvo XC90 for Uber
Herman Herman, a roboticist and commercialization specialist from Carnegie Mellon University’s Robotics Laboratory, currently the director of the facility, thinks that we are long away from autonomous vehicles. Herman has been leading the robotics lab since February 23, 2015, after 40 of its 100 scientists were hired by Uber.

At the time, Uber had a partnership with the Robotics Lab at Carnegie Mellon University, but the ride-sharing giant preferred to hire the specialists it was working with and set up a center in Pittsburgh to deal with self-driving car development. That is the reason why the company will first test autonomous Volvo XC90 SUVs on the streets of that city, and not someplace else.

According to Herman Herman, there is a good reason why Uber will have engineers behind the wheel of its autonomous driving prototypes. Instead of just having a “safety net,” Herman states that the current condition of the technology that means to deliver driverless vehicles is “not even close” to its stated goal.

From a technical perspective, Herman told Vice that he is not sure whether Uber’s technology is ready for self-driving operation. Instead, he believes that the company said otherwise because it needed the marketing praise that goes with driverless vehicle technology. Other non-technical reasons were also involved.

The most important statement in Herman’s interview with Vice’s Motherboard is that it is “easy” to achieve 80% of the development for a driverless vehicle, which consists of operating in areas with excellent weather, good GPS, and proper road markings.

However, the real problem is to attain the rest of the development process, which requires making a car that can drive itself when there is no GPS, inclement weather, unmarked roads, and even all of the above.

While humans have a shot at achieving this and driving in such situations as the one described above, driverless cars of today cannot perform such actions.

That is why the vehicles tested in California performed so well - all they had to endure was bright sunlight or rain. From that point of view, we believe Mr. Herman is right when he says it is a crucial obstacle that needs to be resolved.
If you liked the article, please follow us:  Google News icon Google News Youtube Instagram X (Twitter)
About the author: Sebastian Toma
Sebastian Toma profile photo

Sebastian's love for cars began at a young age. Little did he know that a career would emerge from this passion (and that it would not, sadly, involve being a professional racecar driver). In over fourteen years, he got behind the wheel of several hundred vehicles and in the offices of the most important car publications in his homeland.
Full profile

 

Would you like AUTOEVOLUTION to send you notifications?

You will only receive our top stories