Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

BMW will be autonomous in 2021 how about Tesla?

This site may earn commission on affiliate links.
I know Elon has talked a bit in the past about autonomous cars, and as I recall, he's mentioned the 2018 timeframe. Now that BMW has planted a stake in year 2021 for autonomous, I hope Tesla comes out with a more definitive date for their autonomous system.

I, Robot car: BMW to launch autonomous electric flagship in 2021 - Roadshow

Ideally I'd like to hear that the M3 will come with the hardware for autonomous level 4, but I suspect the software and laws may need additional time. If the initial M3s ship with the needed hardware, I will likely outright buy the M3. If autonomous driving is missing from the initial launch of the M3, I'll probably lease for 3 years and by the time the lease is up autonomous should be either out or much closer to reality.

I saw in the MS forum that there's likely AP v.Next coming to the MS sometime this year due to leaked schematics. I'm by no means an autonomous expert, but my gut tells me even those AP enhancements aren't yet up to the level needed for full autonomy. I would think more cameras and sensors are needed for full level 4.
 
  • Funny
Reactions: jelloslug
That 2021 date sounds like a lot many other things are predicted to happen in 5 years' time, but constantly gets pushed out another 5 years once the previous date had passed: A cure for cancer in 5 years; cold fusion in 5 years; hydrogen fuel cell cars will be practical and widely available in 5 years, etc.

I'm optimistic that fully autonomous vehicles will come to fruition some day, but I'm not going to guess whether it'll be in 5, 10 or 20 years. I hope it'll be in my lifetime so I can retain my mobility and freedom even after I'm too old to drive myself. I don't follow BMW, but do they currently offer any type of driver-assist similar to Tesla's autopilot in their current lines of cars? Or is all their work still in the testing phase? If it's just the latter, it seems like an awfully big jump from zero to fully autonomous in 5 years, unless someone like Google will license/partner their technology to BMW and other manufacturers.

If I were to guess, I'd say the first iteration of Model 3's will not include the full complement of hardware needed for fully autonomous driving.
 
The best reason for the 5 year date is that the masses forget almost everything in 5 years.

Google is spear heading the changes needed for level 4 full antonymous.
The government office that certifies the antonymous levels require a human be in charge of driving so that kind of messes up fully antonymous.
 
Sounds like they build "backup systems" - I'm thinking for example secondary hydraulic assistance for braking or steering when electrical primary systems fail. Not necessarily the autopilot itself, but systems that are integrated with it or serve as emergency failsafe for it.
 
  • Informative
Reactions: GoTslaGo
I don't care who is first to a fully autonomous car, because my focus is going to be on the safety of it. Tesla is making it useful long before anyone else (even though others have similar features, they aren't nearly as useful IMO) and so I'm inclined to trust them.
 
The best reason for the 5 year date is that the masses forget almost everything in 5 years.

Google is spear heading the changes needed for level 4 full antonymous.

Is that Level -4?

The government office that certifies the antonymous levels require a human be in charge of driving so that kind of messes up fully antonymous.

Well, if fully autonomous has the car do everything, then for fully antonymous you'd probably require a human to be in charge of everything. ;)
 
Tesla has chosen Korean company Mando Corporation to help it develop a key component of autonomous driving technology.
Read more at Tesla chooses Korean partner Mando for self-driving technology.


I thought Mobileye was in the front of the Autopilot pack, especially with Tesla. Then I see this Mando thing. Am I mixing apples/oranges with auto pilot and auto drive?
Story has been updated to say:
Updated May 23 4:45pm: We received word from a Tesla communications rep indicating that Mando does not supply autonomous technology systems and the original story published by the Korea Herald is untrue. We were told that Tesla currently sources steering racks for its Model S from Mando.
 
Last edited:
  • Informative
Reactions: GoTslaGo
I know Elon has talked a bit in the past about autonomous cars, and as I recall, he's mentioned the 2018 timeframe. Now that BMW has planted a stake in year 2021 for autonomous, I hope Tesla comes out with a more definitive date for their autonomous system.

You're new around these parts aren't ya? Elon never lets himself be defined by any dates - stuff just shows up. And Tesla is WAY ahead of BMW in the race for autonomy. The current crowd poll in the Model S forum has most votes clustering around an 8-14 month timeframe for a multi-camera hardware release in the Model S/X. Go read those posts and watch the Mobileye presentation videos - will be very helpful for you.

Poll: Testing the wisdom of crowds - how far off is Autopilot V2 Hardware?
 
  • Like
Reactions: GoTslaGo
That 2021 date sounds like a lot many other things are predicted to happen in 5 years' time, but constantly gets pushed out another 5 years once the previous date had passed ...If I were to guess, I'd say the first iteration of Model 3's will not include the full complement of hardware needed for fully autonomous driving.

As someone with an autopilot Model S I can confirm we are much closer to autonomy than most of the public thinks we are. The Tesla system with a single camera is astounding in terms of how bad the roads can be and yet the car still stays in its lane. Go read the post above this one and the Mobileye presentations. It also has been getting steadily more accurate over the 6 months since its release in successive new firmware releases.

All signs point to full-autonomy-capable hardware in the Model S 12 months or less from now - that is based on 360 degree camera coverage supplemented by radar + sonar, using 5 linked EyeQ3 SOC's for processing - later versions using a single EyeQ4. Software of course will take more time.

Read the notes and watch the latest Mobileye presentations. Read the notes on their work in building sub-components of machine learning specifically trained to deal with human drivers (scenarios such as round-abouts) and about Mobileye's crowd-sourced, camera-based, low-bandwidth-needed high precision mapping project launching shortly.
 
  • Like
Reactions: GoTslaGo
Until each of the states do something about poorly marked roads autonomous driving will not be possible. Who will pay for this?
Because this is an impossible circumstance (even if every road was painted today, tomorrow there will be a repair, new segment, gap) Trying to get every state (and country) to paint every road is not a logical solution to self-drive. Lets not even talk about who pays- can't be done this way.
What CAN be done is develop a learning system, where a driven car photographs and reports to the Borg what it found, then the next car will know that the curb is gone. Then, cameras may recognize the edge of a road, even unpainted. Then it connects the dots between paint segments. Then dirt roads get recognized- edges are noted, and a center position in your lane may only need one edge.

I see software and camera being helpful, not painting crews. And not this week, but maybe in the next few years.
 
I don't care who is first to a fully autonomous car, because my focus is going to be on the safety of it. Tesla is making it useful long before anyone else (even though others have similar features, they aren't nearly as useful IMO) and so I'm inclined to trust them.
Talking about safety. I may be more calm to drive myself than be watching constantly for the autonomous driving mistake. Or is that not for most?
 
  • Like
Reactions: BluestarE3