Driverless cars... how far off are they now?

Forums › Transport (cars, bikes and boats) › Driverless cars... how far off are they now?

1 | 2 | 3 | 4 | 5 | 6 | 7 | ... | 9

675 posts

Ultimate Geek

#1951833 5-Feb-2018 10:21

networkn:

langi27:

maybe some blind optimism on my part but, i still think there is a place for these types of vehicles and they will be here sooner than you think. They may not be in your face on a daily basis but they will operate behind the scenes. think about large companies that uses drivers for specific duties, eg Medical couriers that pick up samples for testing and take them to a testing facility, or a coffee roasting house that sends out its fresh beans to cafes every day. Why can't this be done autonomously?

So long as you agree to walk out to the car drop it off and collect it, I don't see how this wouldn't be cheaper than a man in a van.

Building a car that doesn't need an interior will be significantly cheaper than a new vehicle and probably on par with a 2nd hand jap import. You can already buy cheap Chinese brand new cars for $15k, (Cherry J3) I very much expect a driver less, passenger less car to be half that.

There is blind optimism, and then there is this. :)

Thanks for the irrelevant comment.

ripdog

548 posts

Ultimate Geek
Inactive user

#1951835 5-Feb-2018 10:24

tdgeek:

ripdog:

On motorways, self driving already exists: Teslas, and I think BMWs.

The Waymo cars already work in 99% of scenarios. Hell, even more than that. The work right now is the really hard work getting the tiny tiny edge cases filed down. Ensuring the cars react well to realllly weird and unusual scenarios, and getting disengages down (scenarios where the car flips control to the driver due to it not knowing what to do, or experiencing software or hardware failure).

Of course, the big problem with Waymo cars is that they need an extremely well mapped area to function. Detailed 3d scans of every road they operate on, in order to allow them to easily distinguish between expected and unexpected objects. The maps also require manual placement of things like traffic lights. If a road has been significantly modified from the mapped version, the car will disengage and inform the mothership, so other cars will avoid that road until it's remapped.

Obviously the mapping requirement must be lifted before the car can become a worthwhile mass market product, but who knows how long that will take. Other, more reckless companies (like uber and tesla) may barrel ahead without taking all the time and care that waymo is.

Anyone interested in self-driving cars should read waymo's recently published safety report, where the go over all the work which goes into making the waymo cars safe. It's fascinating stuff: waymo-safety-report-2017.pdf

It seems waymo has started to market the technology, starting with talking about safety, and educating the public about how they work and how they will fit into society, i.e. working with emergency services. Could be pointing at the beginning of a ramp up to launch, perhaps?

Good points but the topic is driverless (autonomous), not self driving.

Huh? Self-driving IS driverless.

I mean, sure, waymo test cars have safety drivers behind the wheel just in case, but the goal is fully driverless cars - and I think the pilot program already offers that for the lucky few who have been accepted.

wellygary

8315 posts

Uber Geek

#1951843 5-Feb-2018 10:34

ripdog:

Huh? Self-driving IS driverless.

I mean, sure, waymo test cars have safety drivers behind the wheel just in case, but the goal is fully driverless cars - and I think the pilot program already offers that for the lucky few who have been accepted.

I think they mean the ability to summon your vehicle from across town, or for it to return home once it has dropped you off...

tdgeek

29746 posts

Uber Geek

Trusted

Lifetime subscriber

#1951851 5-Feb-2018 10:43

ripdog:

tdgeek:

ripdog:

On motorways, self driving already exists: Teslas, and I think BMWs.

The Waymo cars already work in 99% of scenarios. Hell, even more than that. The work right now is the really hard work getting the tiny tiny edge cases filed down. Ensuring the cars react well to realllly weird and unusual scenarios, and getting disengages down (scenarios where the car flips control to the driver due to it not knowing what to do, or experiencing software or hardware failure).

Of course, the big problem with Waymo cars is that they need an extremely well mapped area to function. Detailed 3d scans of every road they operate on, in order to allow them to easily distinguish between expected and unexpected objects. The maps also require manual placement of things like traffic lights. If a road has been significantly modified from the mapped version, the car will disengage and inform the mothership, so other cars will avoid that road until it's remapped.

Obviously the mapping requirement must be lifted before the car can become a worthwhile mass market product, but who knows how long that will take. Other, more reckless companies (like uber and tesla) may barrel ahead without taking all the time and care that waymo is.

Anyone interested in self-driving cars should read waymo's recently published safety report, where the go over all the work which goes into making the waymo cars safe. It's fascinating stuff: waymo-safety-report-2017.pdf

It seems waymo has started to market the technology, starting with talking about safety, and educating the public about how they work and how they will fit into society, i.e. working with emergency services. Could be pointing at the beginning of a ramp up to launch, perhaps?

Good points but the topic is driverless (autonomous), not self driving.

Huh? Self-driving IS driverless.

I mean, sure, waymo test cars have safety drivers behind the wheel just in case, but the goal is fully driverless cars - and I think the pilot program already offers that for the lucky few who have been accepted.

There is a significant difference between driverless (no driver present) and self drive (driver present)

ripdog

548 posts

Ultimate Geek
Inactive user

#1951854 5-Feb-2018 10:44

wellygary:

ripdog:

Huh? Self-driving IS driverless.

I mean, sure, waymo test cars have safety drivers behind the wheel just in case, but the goal is fully driverless cars - and I think the pilot program already offers that for the lucky few who have been accepted.

I think they mean the ability to summon your vehicle from across town, or for it to return home once it has dropped you off...

That's the icing on the cake of the self-driving tech. Once the car can drive itself, all that cool other uses like super-cheap taxis will pop up out of nowhere. I think waymo is planning on doing their own such service.

And no doubt your waymo car will indeed drive itself home if you order it to, though I have no idea why you'd want it to.

MikeB4

18435 posts

Uber Geek

ID Verified

Trusted

#1951856 5-Feb-2018 10:46

I see driverless cars everywhere.

langi27

675 posts

Ultimate Geek

#1951868 5-Feb-2018 10:54

there are lots of reasons you'd want to send your car home, (or forward to another destination)

send your car home to charge up, like your robotic vacuum cleaner.

Saves you finding and paying for a park,

Probably save on insurance as its being stored at home rather than on the street

Drops kids off at school then comes back for someone else

Great for when you forget something and can just send the car back to collect (so long as someone there to put it in the car).

Drops you off at work before going to get itself serviced

Lenovo

Shop now for Lenovo laptops and other devices (affiliate link).

tdgeek

29746 posts

Uber Geek

Trusted

Lifetime subscriber

#1951879 5-Feb-2018 11:05

langi27:

there are lots of reasons you'd want to send your car home, (or forward to another destination)

send your car home to charge up, like your robotic vacuum cleaner.

Saves you finding and paying for a park,

Probably save on insurance as its being stored at home rather than on the street

Drops kids off at school then comes back for someone else

Great for when you forget something and can just send the car back to collect (so long as someone there to put it in the car).

Drops you off at work before going to get itself serviced

Yep

Went to pub, had a couple too many, send it home.

Due for service, send to dealership, they can then send it home.

When it is fully reliable and as foolproof as it can be, it would be a great time and/or money saver. Brick and mortar stores would include a "put inside car" real time delivery service

kryptonjohn

2523 posts

Uber Geek

Lifetime subscriber

#1951894 5-Feb-2018 11:37

ripdog:

wellygary:

ripdog:

Huh? Self-driving IS driverless.

I mean, sure, waymo test cars have safety drivers behind the wheel just in case, but the goal is fully driverless cars - and I think the pilot program already offers that for the lucky few who have been accepted.

I think they mean the ability to summon your vehicle from across town, or for it to return home once it has dropped you off...

That's the icing on the cake of the self-driving tech. Once the car can drive itself, all that cool other uses like super-cheap taxis will pop up out of nowhere. I think waymo is planning on doing their own such service.

And no doubt your waymo car will indeed drive itself home if you order it to, though I have no idea why you'd want it to.

Yep. On the other hand, the technology for self drive is 90% of the technology required for driverless, isn't it?

Check it out: https://www.youtube.com/watch?v=wuhbqcMzOaw

Suspect one the biggest problems with introducing driverless cars are the legacy cars (specifically their drivers) sharing the road. Might help if the legacy cars all had some sort of transponder so the self drive cars could have a better clue of what the hell they're up to.

jpoc

1043 posts

Uber Geek

#1951899 5-Feb-2018 11:45

How far off?

I have some news for all of the folks here saying stuff like 'not in my lifetime'.

Next year.

GM will launch a range of driverless cars in 2019.

https://www.wired.com/story/gm-cruise-self-driving-car-launch-2019/

No steering wheel, no pedals.

They are already tooling up the manufacturing facility.

Ford and Volvo have been scheduling 2020-2021 for their own driverless cars and they have already signed contracts with the likes of Uber and Lyft.

GM will be launching their own competitor in that space. They are doing that because they recognise that, in the long term, the profits in the automoble industry will go to the operators of the 'Transport as a Service' businesses and not the car makers.

tdgeek

29746 posts

Uber Geek

Trusted

Lifetime subscriber

#1951948 5-Feb-2018 12:02

jpoc:

How far off?

I have some news for all of the folks here saying stuff like 'not in my lifetime'.

Next year.

GM will launch a range of driverless cars in 2019.

https://www.wired.com/story/gm-cruise-self-driving-car-launch-2019/

No steering wheel, no pedals.

They are already tooling up the manufacturing facility.

Ford and Volvo have been scheduling 2020-2021 for their own driverless cars and they have already signed contracts with the likes of Uber and Lyft.

GM will be launching their own competitor in that space. They are doing that because they recognise that, in the long term, the profits in the automoble industry will go to the operators of the 'Transport as a Service' businesses and not the car makers.

Great stuff. But, what Govt will allow these to hit our roads? When they are FULLY tested they will, but I cant see that being simply a new VINZ form to fill in. The last thread on autonomous cars highlighted all the issues. From death to who pays when my auto car hit yours? Great stuff but unlike smartphones that have bugs, we cant afford versions 1.0 and 2.0 on the roads

kryptonjohn

2523 posts

Uber Geek

Lifetime subscriber

#1951950 5-Feb-2018 12:04

jpoc:

How far off?

I have some news for all of the folks here saying stuff like 'not in my lifetime'.

Next year.

GM will launch a range of driverless cars in 2019.

https://www.wired.com/story/gm-cruise-self-driving-car-launch-2019/

No steering wheel, no pedals.

They are already tooling up the manufacturing facility.

Ford and Volvo have been scheduling 2020-2021 for their own driverless cars and they have already signed contracts with the likes of Uber and Lyft.

GM will be launching their own competitor in that space. They are doing that because they recognise that, in the long term, the profits in the automoble industry will go to the operators of the 'Transport as a Service' businesses and not the car makers.

Cool - but I'll believe it when I see it. Marketing slicks in pinstripe suits will say anything!

Dairyxox

1594 posts

Uber Geek

#1951954 5-Feb-2018 12:14

Seems likes its going to be 3-5 years off for the foreseeable future.

It appears the the Dunning-Kruger effect applies to those who attempt the task, and we're now beginning to see the 1st downward slope.

tdgeek

29746 posts

Uber Geek

Trusted

Lifetime subscriber

#1951957 5-Feb-2018 12:25

The relevant discussion is what needs to happen for this be a reality?

1. Proximity awareness of the car. Not hard but infrastructure needs to be created or modified to give the car full awareness of roads, traffic, parked cars, fixed items, people, animals, traffic lights.

2. Legal/insurance decisions over accidents.

All these are todays tech, but it needs to be standardised, then developed from what we already have, then tested, then implemented

gzt

17111 posts

Uber Geek

Lifetime subscriber

#1951962 5-Feb-2018 12:43

langi27: They may not be in your face on a daily basis but they will operate behind the scenes. think about large companies that uses drivers for specific duties, eg Medical couriers that pick up samples for testing and take them to a testing facility, or a coffee roasting house that sends out its fresh beans to cafes every day. Why can't this be done autonomously?

Flying drones are more likely than autonomous wheels for some applications.

1 | 2 | 3 | 4 | 5 | 6 | 7 | ... | 9