Tesla BEV Competition Developments

[Auzie]

A Mercedes S-Class BEV? Can't imagine what that would cost! Also mentioned is a plug-in hybrid with 80kwh battery with only 19 miles or range, huh?

Three Mercedes Plug-In Hybrids On Sale By End Of 2015: S-Class, C-Class, ML-Class

Plenty of hybrids on the way. The article mentions "a notion of" Merc S C class bev in 2020. My interpretation: there is no plan in MB camp for bev before 2020.

By 2020, it will be hard to catch up with Tesla.

 

[CHGolferJim]

I imagine by 2020 the relationship between Tesla and MB will have greatly evolved, especially if Elon decides to cash out to live on Mars. Seems analogous to a Big Pharma company toying with a biotech research partner.

 

[RobStark]

I imagine by 2020 the relationship between Tesla and MB will have greatly evolved, especially if Elon decides to cash out to live on Mars. Seems analogous to a Big Pharma company toying with a biotech research partner.

Except in this case the little research partner has a market cap 1/3 of the established player.

And the little research company has an established retail brand name and a small but growing distribution network.

I doubt SpaceX will have a colony on Mars by 2020.

Elon hopes to live on Mars after he reaches retirement age. Not 50.

He has said more than once he will be the last to sell his TSLA shares.

 

[SebastianR]

Except in this case the little research partner has a market cap 1/3 of the established player.

And the little research company has an established retail brand name and a small but growing distribution network.

Yep, as we have seen, the automotive industry is by far not as stable as people may think - as most ICE car makers today are essentially owning design, sales/marketing and the engine department and as the later one is getting obsolete rather quickly, I'm not sure at all if MB will be what it is today in 20 years.

I can't wait for all the existing ICE car makers to be disrupted - it really is time now.

 

[Auzie]

Google is not complacent with just software. They’re making big investments into hardware (ie., Nest). And they will likely continue to invest into hardware.

Previously the google self-driving car was pretty much all software and some hardware (i.e., Lidar cameras, etc). But with this new version of the google self-driving car, we see that Google is designing the entire car not just the software. They’ve designed the powertrain, the body, the hardware, the software. It’s the complete package.

They can easily hire a manufacturing partner (ie., like Apple hires Foxconn to make iPhones) to make the vehicle and Google can sell them to cities, companies, etc.

Google will roll this out very slowly and very thoroughly, meaning that if they go for Manhattan then they will do a thorough mapping/scoping of every square foot of drivable pavement in the city before rolling it out in Manhattan.

———————

Some final thoughts:

1. To understand the google self-driving car, one needs to look out 10-20 years. As long as Google iterates at the pace and commitment that they’ve shown, this technology is a major disrupting force for dense urban transport. The initial launch implementation will likely take longer than most people expect (ie., maybe 5 years to get it on the road). But most people are likely underestimating the truly disruptive nature of what Google is showing. In 15 years, it’s possible that the densest urban cities could be filled with “Pods” like this to take people from point to point. That would be truly revolutionary.

2. Google will have trouble breaking into the regular consumer auto market because it’s so much more difficult to do a truly self-driving car at 45mph+. But if urban cities take to the Google “pod” within 15 years, then this could give Google a lot of momentum to branch into the consumer auto market as their self-driving tech improves. In 15-20 years, Google and Tesla could be the biggest names in auto.

Jalopnik claims that Roush is helping build Google car. I think this is smart business move by Google, outsource difficult bits to someone else and stick with what Google does best.

I agree with your final thoughts. The self driving control system for cars is highly disruptive. It might be difficult to see a business case for Google car now, however it was equally difficult to see a business case for the search engine decades ago. The revolutionary disruption in case of success may be sufficient to justify initial high investment.

If Google manages to develop their system to be safe at higher speeds, I will be extremely happy both as a Google investor and as a driver.

There is some sort of competition between Tesla and Google on this front, namely Elon claimed that Google approach with radars, lasers and cameras is too expensive to be commercialized. Tesla's approach is to use less expensive cameras to develop auto guidance system with driver override.

In Elon's words: Approach is 360 deg flush mounted tiny cameras + radar (prob not lidar). Lot of software & hardware level image processing. In Elon's words,"My opinion is it's a bridge too far to go to fully autonomous cars," Musk explains. "It's incredibly hard to get the last few percent." The executive says that he envisions self-driving technology operating much like an auto-pilot system that drivers could turn on in most situations."

I tend to agree with Elon, however anything is possible, Google will start testing their car soon. Google has much less on the line compared to Tesla due to its cash generating ability and position.

Few interesting bits and pieces on Google car:

1. Fully electric
2. Vision up to 180m away
3. No controls
4. 25mph max speed for now

 

[green1]

I don't think higher speed is really the problem, it's unusual situations. Computers just aren't that good at dealing with the unknown, and I don't trust that the self driving car will do well in conditions that can trick a camera image (snow obscuring lane markings, wet roads where is easier to see where the lane making used to be then where they are now, damaged street sign, etc.) or where the usual inputs are replaced by something unexpected (police officer waving you in a certain direction while a sign tells you to do something else, or a row of pylons that's set up a little funny, etc)

There's a reason online forms use captchas to verify humans separate from computers, computers aren't good at image recognition. I think Elon is right on here, these technologies make great driver assists, but I think they're still a long way from full driver replacement in all situations.that last few percent, the edge cases, will be a nightmare.

 

[Auzie]

I don't think higher speed is really the problem, it's unusual situations. Computers just aren't that good at dealing with the unknown, and I don't trust that the self driving car will do well in conditions that can trick a camera image (snow obscuring lane markings, wet roads where is easier to see where the lane making used to be then where they are now, damaged street sign, etc.) or where the usual inputs are replaced by something unexpected (police officer waving you in a certain direction while a sign tells you to do something else, or a row of pylons that's set up a little funny, etc)

There's a reason online forms use captchas to verify humans separate from computers, computers aren't good at image recognition. I think Elon is right on here, these technologies make great driver assists, but I think they're still a long way from full driver replacement in all situations.that last few percent, the edge cases, will be a nightmare.

I agree it is not higher speed, speed is limited for perceptions of safety during trials.

Elon's approach may be right for the moment. Tesla has a chance to win the race to be the first to make commercial driver assist systems.

It might be difficult to see Google succeeding but I keep an open mind for their approach as well. Current technology far surpassed my past ability to imagine, so I expect future technology to be beyond my current grasp.

I am quite excited to follow this race between Tesla and Google. Google does have some advantages imo, they steal Tesla engineers, they have much more cash, very stable business, much less business risk to deal with, and may be better positioned in this race. On the other side, Tesla has Elon and that might just prevail.

 

[erha]

I actually think both Tesla and Google are right in their own ways. Google's advanced system might be too expensive to install in every car sold. However, I don't think that's their intention. I think they want to develop a taxi fleet of autonomous cars, which means that every single car can drive 24 hours a day and in one way replace several household cars. That way it doesn't really matter if the cost of the car is twice that of a car that is 95 % autonomous. I think that is how transportation will be in the future (beginning in urban areas with lack of parking spots), but this is not going to happen in a few years. Tesla's approach is in my view more likely to succeed first and then I think both Google and Tesla will move into the autonomous taxi fleet business.

 

[RobStark]

Google won't replace several household cars but taxi cabs with human drivers. Some will switch from trains and buses.

$4 plus per mile to $0.50 per mile. Average cars that drive 12k miles won't get replaced. Even ones with 8k miles per year.

SF is the obvious first target then NYC. Then other densely populated Northeastern cities.

 

[Robert.Boston]

Google won't replace several household cars but taxi cabs with human drivers. Some will switch from trains and buses.

$4 plus per mile to $0.50 per mile. Average cars that drive 12k miles won't get replaced. Even ones with 8k miles per year.

SF is the obvious first target then NYC. Then other densely populated Northeastern cities.

Think about ZipCars, which have successfully displaced many people's demand for a(nother) car. Its success has been mostly in urban/campus areas where people can easily walk to a central depot. But with self-driving cars, ZipCar and its competitors can expand to serve lower-density markets: the car will come to your door. Considering the number of people with 2+ cars in the household, it's probably easy to reduce their average car holdings with such a program. (And, to finish the thought, that means the autonomous car can be more expensive because its cost is being shared among many households.)

 

[jerry33]

I don't think higher speed is really the problem, it's unusual situations. Computers just aren't that good at dealing with the unknown, and I don't trust that the self driving car will do well in conditions that can trick a camera image (snow obscuring lane markings, wet roads where is easier to see where the lane making used to be then where they are now, damaged street sign, etc.) or where the usual inputs are replaced by something unexpected (police officer waving you in a certain direction while a sign tells you to do something else, or a row of pylons that's set up a little funny, etc)

You're right but there are also some situations where cameras would be better. I'm thinking ice fog. Those from Saskatchewan and Manitoba can relate. You're driving along and wham, it's like the hood opened. You have to open the driver's door to see the paint lines. IR or UV cameras would help here. Also from twilight to dawn where the eye doesn't see as well as the camera.

The big question is where are the pull down screens shown in Dragon 2?

 

[green1]

That's exactly my point though, these are excellent assist systems, all this technology can do a great job helping the driver do their job better in cases where the driver would be otherwise at a disadvantage. But that's a long way from a full replacement in all situations. I honestly do believe we will get to self driving cars. I just don't think it's as close as the proponents want us to believe.

 

[DaveT]

I don't think higher speed is really the problem, it's unusual situations. Computers just aren't that good at dealing with the unknown, and I don't trust that the self driving car will do well in conditions that can trick a camera image (snow obscuring lane markings, wet roads where is easier to see where the lane making used to be then where they are now, damaged street sign, etc.) or where the usual inputs are replaced by something unexpected (police officer waving you in a certain direction while a sign tells you to do something else, or a row of pylons that's set up a little funny, etc)

There's a reason online forms use captchas to verify humans separate from computers, computers aren't good at image recognition. I think Elon is right on here, these technologies make great driver assists, but I think they're still a long way from full driver replacement in all situations.that last few percent, the edge cases, will be a nightmare.

It's the tricky situations that the Google self-driving car will have to master in order of the technology to get adopted. The advantage though of Google's approach is that they're doing it at low speeds, ie., 25mph (maybe 35mph in 5-10 years). Coming to a stop from going 25mph is relatively easy (compared with coming to a stop from 65mph). So, if the self-driving car spots a situation it can't handle then it can easily slow down (ie., to 5-10mph) and even ask the driver for info. It could also just come to a complete stop if needed (since it will be driving only on streets with a 35mph speed limit). If a self-driving car was going on the freeway, there's no way it could stop or significantly slow down on the freeway. That's one of the major reasons why getting self-driving right at higher speeds is so difficult. In other words, doing it at lower speeds gives you a lot more time for the car to react and the car can slow down or even stop if needed. The faster the car is going the more difficult it becomes to stop suddenly in case of emergencies.

 

[pz1975]

I am curious if a self-driving car could be designed to not just avoiding hitting things, but also to avoid being hit using anticipation and avoidance movements.

For example, one potential situation for this is driver error from behind, although this is probably only relevant at freeway speeds. When there is a sudden stop/slow down up ahead and you have slowed down enough that you know you won't hit the car ahead of you, occasionally you will see the car in the rearview coming in too fast and likely will rear end you. The proper move of course is to quickly pull over to the side (assuming it is a shoulder or no one is there) and give him/her that extra room to stop (or at least hit someone else!).

I would be blown away if a self-driving car could 'learn' to do something like that.

 

[DaveT]

It's the tricky situations that the Google self-driving car will have to master in order of the technology to get adopted. The advantage though of Google's approach is that they're doing it at low speeds, ie., 25mph (maybe 35mph in 5-10 years). Coming to a stop from going 25mph is relatively easy (compared with coming to a stop from 65mph). So, if the self-driving car spots a situation it can't handle then it can easily slow down (ie., to 5-10mph) and even ask the driver for info. It could also just come to a complete stop if needed (since it will be driving only on streets with a 35mph speed limit). If a self-driving car was going on the freeway, there's no way it could stop or significantly slow down on the freeway. That's one of the major reasons why getting self-driving right at higher speeds is so difficult. In other words, doing it at lower speeds gives you a lot more time for the car to react and the car can slow down or even stop if needed. The faster the car is going the more difficult it becomes to stop suddenly in case of emergencies.

Just read this interesting article on google's self driving tech. This is from a few weeks ago, seemingly prior to their latest reveal.

The Trick That Makes Googles Self-Driving Cars Work - Atlantic Mobile

 

[Auzie]

Electric but not bragging about it

Merc B Class with Tesla powertrain is Mercedes Benz first battery electric car.

In my view there is no bragging because there seems to be nothing to brag about.

Range 85 miles, range plus 100 miles (battery charged more, to 31.5kWh instead of standard 28kWh).

$42,375 price

Economy and Sport mode (177hp for sport mode)

Reasoning that baffles my mind, quote from the article: “A lot of Silicon Valley people don’t want to scream, ‘Hey, I’m driving an electric car,’ ” Mark A. Webster, Mercedes-Benz USA’s general manager for e-mobility, said as we spun through Palo Alto. “They care about the environment, but they don’t have to brag about their sustainability consciousness.”For all the excesses of this high-flying capital of software, not every well-heeled executive wants a big, flashy and fast electric flagship or a small futuristic E.V. When the B-Class electric goes on sale this summer in select states, mostly on the coasts, they will have a new choice.

My interpretation: MB chose to underdesign their first bev with the nonsensical excuse. MB is dragging their feet. The statement above makes no sense to me.

 

[ecarfan]

I've read everything I can find about this new MB EV and it's hard to get excited about it. Looks like a shrunken mini van, range is modest, acceleration so so, interior styling boring. The only pluses to me are that it has a Tesla powertrain and the gear selector stalk works like the S so it would be familiar. Maybe a test drive would change my mind.

 

[Auzie]

I've read everything I can find about this new MB EV and it's hard to get excited about it. Looks like a shrunken mini van, range is modest, acceleration so so, interior styling boring.

Would you consider buying that car?

I ruled it out at the first look, without any further knowledge.

Also I find many flaws in MR. Webster's reasoning, the first being that any well-heeled executive from Silicon Valley would be in the market for that car.

 

[DaveT]

I've read everything I can find about this new MB EV and it's hard to get excited about it. Looks like a shrunken mini van, range is modest, acceleration so so, interior styling boring. The only pluses to me are that it has a Tesla powertrain and the gear selector stalk works like the S so it would be familiar. Maybe a test drive would change my mind.

True, but it might be the best EV under $50k. I know that's not saying much though.

 

[SteveG3]

Would you consider buying that car?

I ruled it out at the first look, without any further knowledge.

Also I find many flaws in MR. Webster's reasoning, the first being that any well-heeled executive from Silicon Valley would be in the market for that car.

I saw the car at the NY auto show. I thought it was reasonably nice looking and the interior was quite nice for a mid $30K car (assuming the full tax rebate). A Mercedes that seats 5 with a cost of ownership comparable to a well-equipped Camry or Accord? This might partly be my Tesla bias, but I think a family with two cars could be quite satisfied with this car and another to do any road trips. All that said, looking at the photo I do see the shrunken mini van (hard not to see it in the picture now ), but it never occurred to me in person.