The race for the development of fully autonomous cars has been one of the most important objectives of different companies for some years now, although not all of them have the same objectives and they do not face this race from the same perspective. However, they all have a clear point of contact, and that is that we could only consider a vehicle as fully autonomous. if it reaches level 4.
The use of artificial intelligence and deep learning represents one of the most important pillars when creating fully autonomous cars, But is it really possible to achieve that goal? This question has been answered in an interesting article shared by Reuters, where we can see that it could end up being something impossible without a “human touch”.
By “human touch” they mean a certain degree of intervention or support of human professionals that they would be ready to respond, if necessary, to situations in which these autonomous cars could end up in trouble. The idea put forward by the article is that of a kind of human supervisor who can remotely control cars if something happens, an idea that is certainly not crazy, since it makes a lot of sense and could open the door to a new mode of «distance driving» focused as a service.
Imperium Drive claims that autonomous driving systems they do not have the same ability to predict and weigh risks, than human beings, and that they cannot resolve these issues at the same speed either. This, coupled with their inability to react correctly to certain situations, is one of the key challenges for the industry to overcome, and is the main reason why fully autonomous cars may still be several decades away.
The accidents that are happening in the testing phases of different autonomous driving systems support this idea, although as we have said, supervision of human beings could help alleviate this problem.
It is not something simple and it would be necessary to assess whether it is profitable for the costs it represents, but its importance is only growing. Chris Borroni-Bird himself, an independent consultant who has worked with advanced autonomous driving programs from firms such as GM and Waymo, has acknowledged that autonomous driving systems they do not have the capacity for perception and prediction of human beings, and that they are not capable of making decisions at the same level.
With all this in mind, the prediction that Tesla launched at the time seems like a pipe dream. They may even launch a “fully autonomous” car, but whether it is safe and capable of driving at a human level is another story.
