O2 cuts ribbon on UK's first commercial driverless car lab where it'll blend satellite and 5G signal to stay on the road Mobile network O2, notable for its occasional outages, has opened the UK's first commercial lab for connected, driverless cars. Coming more than a year after Vodafone and Samsung showed off a driverless car hooning up the famous Goodwood Hillclimb thanks to 5G and a bunch of Galaxy S10 handsets, O2's Darwin Satcom Lab is aimed …
That last thing I can imagine wanting is a driverless car that depends on its connection via any fecking mobile phone provider.
Can you imagine the 'Driving plans' they will come up with?
That's aside from the signal suddenly dropping out at the most inconvenient moments. I can see that connected driverless cars must must have a great deal of appeal for moby companies but not for thank you very much.
I am inclined to think that for major cities and population centres, a central control would be better so that overall traffic management can be incorporated into the system, punch in your destination and the system will produce the best, most efficient route and then take over.
What could go wrong?
What the internet has done for bank thieves, could be directly ported over for car thieves! But it gets better - think of the opportunities for car jacking also. Leave your kid in the back seat while you run into the shop to pick up a few items, and no more car or kids. This could revolutionize criminality as we know it today. Progess - isn't it great.
Besides, in 2020 who wouldn't be in favor of contactless theft.
I understand that you can't spin much hype out of a technology for improving a mobile phone network, but autonomous driving? Really?
It seems the 5G-Hype must be fed so that companies and governments keep pouring money into it.
However, autonomous cars already are killing people if they do all the computing on-board. What good would introducing a comms channel that is susceptible to outages, bad reception and mixed latency bring into the already brighly burning dumpster fire that is autonomous driving?
My understanding, 5G technologies (not necessarily 5G as we would understand it) help solve the problem of having every car on the road communicate with each other as to where it is relative to other cars. It's effectively peer-to-peer style networking integrated with "classical" mobile comms to base infrastructure - so the cars are talking directly to each other for localised info, back to base for global info, or even receiving that info via another nearby car.
While communicating cars over 5G might be a good thing, this is exactly the thing why I detest all this 5G & autonomous cars hype peddling.
Autonomous cars must always be able to drive safely among other cars, pedestrians, etc. even when no one around talks to them. I.e. an autonomous car must be safe even when there are no other autonomous cars around.
This makes all the "5G helps autonomous driving" utterly useless. In a perfect world in which all cars communicate, such a thing might be useful. However, the world is not perfect, and makers of autonomous care should make sure, that their autonomous cars are safe on their own (the comms network could be down - possibly due to some tinfoil hatters' arson, there could be only dumb cars around, insert reason).
However, in the mean time, people are getting killed while companies try to figure out how to make a car drive autonomously. No 5G network will save the driver of the fancy Hype-Mobile if the Hype-Mobile slams head-on into a lane divider, which it did not recognize.
Or should we equip all lane dividers with 5G, so that they can tell the Hype-Mobile about them?
Of course they need to do that, but in (say) motorway driving, it would allow cars to follow closely if they were autonomous and communicating in some manner. It's not to say that non-autonomous cars can't be in the mix. If you can only have the cars communicating with 80% of other traffic, you don't say "fuck it, we'll have 0% if we can't have 100%".
It's also not exclusively how autonomous cars will/can work, but it helps a great deal. Accidents and congestion happen because a car/driver can only react to what he sees around him:
https://imgur.com/gallery/CIhYAiv
In the example above (a single car swerving briefly, causes a complete traffic jam), if the cars further behind get the message that this situation has happened, they can slow slightly (say from 70 to 60) to avert a traffic jam/stopped traffic. Any non-autonomous car in this situation would likely slow to the same speed if all the other traffic was doing the same. It's how and why temporary lane speed limits work - slowing the traffic a mile back to 60mph will allow a traffic jam to clear and not get worse.
"My understanding, 5G technologies (not necessarily 5G as we would understand it) help solve the problem of having every car on the road communicate with each other as to where it is relative to other cars."
That makes a hill climb, a relatively short, well defined course, not on a public road, with no oncoming traffic, and probably no other traffic at all such a convincing place to display it.
The next step is macro programming your car to go to the shops for you (emailing your shopping list to the shop) or, even better, order your shopping online and have someone else deliver it to you.
More time for you at Beer o'clock, while the technology does what it's always promised to do/deliver - which is great until the adverts (that will inevitably appear in the system, just like Smart TVs) mean your car will detour to get some Free Samples You May Like and then runs out of fuel a mile or two from home because of its abysmal range.
"in getting autonomous vehicles on the road and making the UK's transport network greener."
Although over time autonomous verhicles may make road use more efficient (they will be travelling most of the time unlike meatware driven ones), they will initially simply add to the number of vehicles on the roads competing for driving and parking space. You can see this in any city "blessed" with electric scooters and bike-sharing schemes: space on pavements has become a rare commodity.
Amen to that.
Given that a car is typically a hulking beast of metal and plastic that has no problem killing pedestrians or cyclists at a speed of several meters per second, I would expect a driverless car to be able to react to an obstacle in less than a millisecond and not depend on sending a data packet to some remote server somewhere and wait for the latency, the processing time and the latency of the response before getting the message "BREAAAAAKKK !".
Driverless has to mean better-than-human-reflexes. Having a central computer and depending on network coverage pretty much scuppers that from the start in my opinion.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
