I see that the founder of Cruise Automation recently announced that they are ready to mass produce a « self-driving ready » vehicle, along with their partner, GM. This will be produced along the lines of the Chevrolet Bolt platform. Apparently this will have all the ingredients that an Autonomous vehicle will need to function, travel, brake and steer etc. Around fifty vehicles have already been built, mostly for the Fleet market and production is expected to ramp up from here.
Congratulations to them on this achievement, but ithought wishing to sound too churlish, I do feel that this announcement has something of the feeling of a race about it and is more about someone crossing the finishing line, rather than having a true production ready self-driving car.
To his credit, Doug Parks, GM’s VP of Autonomous Technology and Vehicle Execution, did say that « There’s still a lot to be done yet » and there certainly is. I believe these cars will remain in a mechanically and electrically production « self-driving ready » state in the broadest sense for a long time yet, but certainly not in true Autonomous usability.
The easy bit is getting all the required modules on a vehicle to turn it into an autonomous vehicle. This is meat and drink to any auto maker or even amateur enthusiast. These days. Sensors, Cameras, Lidar, Navigation and all the mechanical, steering and braking components, etc., are all readily available, even from the likes of Ebay and Amazon
The hard bit, isn’t just a bit; it is, as we all know, a huge shift. Firstly there are the regulatory problems, from Governments around the world to agree and approve. These alone will grind down the process in true, red tape bureaucratic fashion. Then there is the massive infrastructure requirements, the not insignificant « non-vehicle » side of things. Hugely important and a massive task. In fact we hear much about vehicle development, but in my view, not nearly so much about the technologies and activities involved in what could amount to the rebuilding of the entire infrastructure in many towns and cities.
But beyond all else, the biggest « obstacle » of all, involves tens of thousands of engineers, currently working across the globe for various companies, enterprises, or for themselves. They are all striving to achieve an understanding of what amounts to what has now become to be the very heart of the entire Autonomous project, Artificial Intelligence. They are all working to get the required level, understanding and deployment of AI to an acceptable controlled state. This of course is essential, vital, before anyone could ever contemplate any kind of safe, mass introduction of autonomous driving vehicles.
But « when », is the big question…
As the American AI researcher and writer, Eliezer Yudkowsky said:
“By far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.”
<<< Back to the Driverless Blog
Could AI be the key to true full autonomous driving?
