Despite tremendous strides being made by Uber, Google and the like in driverless technology, arguably the most interesting angle on this is the advances being made in technology that turns existing vehicles into driverless ones.
I’ve written a few times about companies such as Otto and Oxbotica, both of whom are making strides in this direction. Indeed, Otto was recently bought by Uber to support their own march into driverless technology.
First mover
Arguably the most advanced of these startups however is Comma.ai, who have received considerable publicity lately for their claims of offering a automation service for under $1,000, although there will also be a $24 monthly subscription to pay for software updates. They plan to begin rolling it out by the end of this year.
Now, it should be said that the device won’t make your car completely automated, but it will give it a number of automated features for highway driving, including braking, changing lanes and accelerating.
The device operates via a number of camera sensors that are added to the vehicle. It is reliant however on the radar systems that come in higher end models today, thus meaning it won’t be something that will turn your vintage vehicle into a driverless demon.
The company report that once fitted, it is capable of performing on a par with systems like the Tesla Autopilot. It’s certainly an interesting development, and notable for the speed with which they have managed to come to market, and they are bolshie about their chances of success.
What is noticeable about all of these companies is how they are largely operating outside of major automobile manufacturers, and it remains to be seen how effective they will be without that level of engineering heft behind them. With the product set to hit the shelves in a matter of months, we won’t have to wait long to find out.
If this is true then I believe most of people will have driverless cars for them. However it seems difficult to make a car driverless within 1,000 USD.
An interesting thought, but I hope nobody buys 'driverless' features based on price. Extreme caution is warranted for obvious reasons.
Doesn't look like it's going to happen any time soon.
https://www.scribd.com/document/329218929/2016-10…
Looks like it was in response to their Oct 20th blog post. That blog post was in response to CA DMV inquires in an attempt to avoid regulation. You can't avoid regulation.
NHTSA is not going to accept "we told the driver they had to keep paying attention" as acceptable. Seems perfectly fair to me.
It's a shame that they couldn't prove themselves to the regulator, but understandable that the technology is subject to scrutiny.
The reason i distrust this product is because it hasn't undergone the rigorous testing that the bigger companies have been putting their products through. look at google's self driving car tech. they've logging millions of miles and they still don't feel it's perfect enough to release.
A fundamental problem with unrestricted AI is that its behavior is impossible to fully predict. Humans don't usually have this problem because our intelligence is constrained by our survival instinct; AI has no such constraint.
So have fun killing people in China where the government doesn't care, asshole. As dangerous as the United States can be, I at least don't have to worry about your shoddily-programmed robot cars running me over because their AI felt like it.
You'd have thought he would have got clearance for all of this before the hype machine started up.
Musk has $4 billion in revenue and a team of lawyers to deal with regulations and clearances — comma.ai had two hackers and an intern. No wonder they couldn't handle the pressure.
Shame, though. I wish they released a product, at least, before being hit with the regulatory hammer.