Autonomous Vehicles

Auto parts suppliers are taking action to position themselves in the emerging world of AVs. I read with interest that this week, South Korea's LG Group has acquired ZKW, the Austrian car light specialist, to become the world leader in lighting self driving vehicles. The joint company will work together in developing intelligent lighting solutions with sensors, cameras and communication systems.

What are some of the things ZKW is planning in lighting self driving vehicles? According to Transport Topics:

"
For example, ZKW envisions technology on its website that would detect pedestrians by the side of the road. The car would stop automatically and project a zebra crossing onto the street.

ZKW also is thinking about lighting signals for safely managing groups of automated trucks that would drive in tight convoys."

http://www.ttnews.com/articles/lg-a...eover-focus-shifts-lighting-self-driving-cars

I am impressed with some of the work and thinking being done at ZKW to prepare us for safer AVs. I enjoyed watching a few of the nine short (25 seconds each) videos on ZKW's official website to see how they use light as a communicating mechanism in the self driving vehicle. Fascinating and showing potential answers to some of the questions we have raised here in the Autonomous Vehicles thread in the Motors section of Overclockers. As a road biker I particularly appreciated some of the potential solutions proposed communicating between self driving vehicle and biker.

https://zkw-group.com/en/home/products-and-solutions/autonomous-driving/
 
Last edited:
Nasher said:
TBH I can't imagine having AV cars is going to be all that practical. All these AVs each carrying 1 person to work? nope. It will only really work with..I dunno..a bus.

But then because there's a bunch of people on-board, you need someone else on it to make sure everyone is behaving themselves and not getting rapey or anything. We'll call him a "bus conductor".
Click to expand...
No there be robots by then to do that bus conductor job....:p:D
 
chaparral said:
No there be robots by then to do that bus conductor job....:p:D
Click to expand...

And if we in Western Europe or the US are not careful, these robots will be made exclusively in China. Note that China has little regards for privacy rights in their digital world. This allows for aggressive, invasive technology, esp when building out Artificial Intelligence and Augmented Reality. Compare and contrast this attitude with we in the West, where privacy concerns are front of mind.

Interesting piece in Automotive News from last week's Beijing Auto Show quoting VW as being very impressed with EV, battery and infotainment systems being made in China. We know that Baidu is a leader in China is AVs. Ford's experiment with Alibaba to build automobile towers (think of it as a car vending machine) giving customers three day test drives based on a check out system using facial recognition technology, is also highlighted. Sales of 102,000 EVs in Q1 put China first globally, up 100% year over year.

China's mandate in developing New-energy vehicles (NEVs) is creating innovation. At the Auto Show, 174 new energy vehicles were introduced, a combination of EVs and plug in hybrids, mostly from local players. A local company called Contemporary Amperex Technology Ltd has a $1.3 billion factory in the works that reportedly would go online as soon as 2020. It could catapult the Chinese supplier to being the world's top producer of lithium ion batteries.

The Alibaba AliOS is considered a world leader in infotainment systems, now in 600,000 internet enabled vehicles. Ford will bring AliOS into its Kuga crossover in China.

http://www.autonews.com/article/20180430/OEM06/180439973/alibaba-baidu-china-auto-technology
 
How pervasive is AI now and why is it such a hot topic?

Wall St Journal (paywall) has an interesting column from Ted Greenwald and a few interesting points that affect self driving vehicles too:

1. What is it and what does it do today?

"It completes half-written words in your texts. It answers questions through your smartphone or smart speaker. It plots the quickest route from here to there. It recognizes familiar faces at your door, curates your social feed, fills your playlist. It slaps ads for those cool boots you considered buying on every webpage you look at."

2. What will it soon do?

"Auto makers are conceiving self-driving vehicles to ferry people to their destinations, then park or carry other passengers until pickup time. Lights and thermostats can connect to your smartphone’s location sensors, so your home can go to sleep when you depart and wake up when you return. Factory robots are sorting goods into customer orders that might be carried to your doorstep by autonomous quadcopters."

3. What's been the pace of change?

"Two decades ago, Deep Blue, the International Business Machines Corp. system that beat world chess champion Garry Kasparov, was the epitome of AI. Now a freebie smartphone game can accomplish much the same thing."

4. What's it good for now?

"In many cases, AI is used to mimic human senses. When you’re searching Facebook photos for pictures of purple berets, the social network’s software assesses shapes and colors and factors in tags entered by users to deliver appropriate snapshots. Likewise, AI is good at recognizing words it hears. It’s excellent at language translation and useful for extracting meaning from written or spoken statements. AI also gives devices a semblance of reasoning ability. The Waze navigation app, for instance, assesses traffic speed to spot jams and predict drive times. The ability to find the most efficient path comes in handy in tasks like optimizing electrical grids and other large networks."

5. How smart is AI?

"AI lately has become far smarter than it once was, but it’s still not terribly smart. Current technology can be fairly sharp-witted when it’s confined to narrow domains, but it still doesn’t cope well with the wide world."

6. How will it change our lives?

"AI is the new electricity,” declared Andrew Ng, chief executive of factory-automation specialist Landing.ai and former head of Baidu’s autonomous-vehicle program and founder of the Google Brain deep learning project. In a speech last year, he said machine learning is poised to transform one industry after another, spurring radically new capabilities and new businesses the same way Thomas Edison’s discovery did beginning in the late 19th century."
 
Summary of comments on AVs from Elon Musk yesterday:

1 Sees evolution of shared electric autonomous vehicles.
2. Sees shared vehicles likely to be a combo of Uber/Lyft and AirBnB.
3. Sees car owner opting in to sharing vehicle and choosing it to be used by third parties at will, restricted to 5-star customers, friends and family only, or not at all.
4. Sees progress at Tesla towards full autonomy with their vehicles capable of a mere AV "upgrade": specifically a computer upgrade to be a neural network, which is an easy "plug in" to their existing fleet.
5. What is holding back their commercial launch? regulation. Currently Musk believes their AVs reduce mortality rate (cites 1.2 million road traffic deaths per annum globablly) now by 50% but somehow the public has it in mind that AVs will only be safe when there are zero fatalities by AVs which Musk sees as impossible.
6. Regulators are responding to public pressure and misleading journalists which he is very upset about because an AV fatality today is front page news whereas no one remembers or reports on the other 1.2 million road traffic deaths with the same intensity.
7. From a technical standpoint, he sees Tesla ready to launch Level 4 and above AVs by the end of 2019.
 
I think Musk is running scared - he doth protest too much, methinks
seems to be trying to enter the propaganda forum to counteract the negative AutoPilot press (fake news?) and get in with the truth, it's like the current anti Russia meme.

The 40% safety improvement he attributes to AP seems controversial
https://www.bloomberg.com/news/arti...tic-in-autopilot-defense-comes-under-scrutiny
it is too commercially sensitive to reveal too much about the systems capabilities, unless there is some legislated disclosure which all vendors have to adhere to.
I would like to see AP1 versus AP2 statistics break out - is his commercial decision to drop Mobileeye compromising safety .. but is he going to reveal that ?


On a different topic of waymo - are there any good articles on the computing techniques being used to make decisions about how the car makes control decisions ?
there is lots of information on deep-learning/neural nets/AI for image identification/decomposition, but for the control decisions it seems much more of a conventional/linear coding strategy.
https://www.theatlantic.com/technol...ret-testing-and-simulation-facilities/537648/
... so how many lines of code, how much data in the car ....
 
sesevans said:
5. What is holding back their commercial launch? regulation. Currently Musk believes their AVs reduce mortality rate (cites 1.2 million road traffic deaths per annum globablly) now by 50% but somehow the public has it in mind that AVs will only be safe when there are zero fatalities by AVs which Musk sees as impossible..
Click to expand...

Well from what we've seen so far, their no safer at all really and they haven't even released all the data :/

At this point the end goal seems to be more about creating a new way to generate money rather than safety.
 
Last edited:
jpaul said:
.....
On a different topic of waymo - are there any good articles on the computing techniques being used to make decisions about how the car makes control decisions ?
there is lots of information on deep-learning/neural nets/AI for image identification/decomposition, but for the control decisions it seems much more of a conventional/linear coding strategy.
https://www.theatlantic.com/technol...ret-testing-and-simulation-facilities/537648/
... so how many lines of code, how much data in the car ....
Click to expand...

As you suggest elsewhere, for competitive reasons the answer is: not that I have seen. After the recent experience with Levandowski, Uber and the lawsuit, I think the amount of info from the "black box" will be rather limited. They will reveal more when it suits them to do so.

Thanks for reminding me to re-read that excellent link you provide to the Atlantic article.

Will we see any significant announcements from Waymo during Google I/O next week?
 
Nasher said:
Well from what we've seen so far, their no safer at all really and they haven't even released all the data :/

At this point the end goal seems to be more about creating a new way to generate money rather than safety.
Click to expand...

The fact that not enough data has been released to confirm safety, it would be wrong to assume the data is there but not released. For competitive reasons, I assume most AV company testing is kept secret at this point. Tesla says they are going to release more data as my next post describes.

As to generating money, why should public companies be penalised for generating profits, esp when the goal of reducing road traffic deaths is a worthy one?
 
jpaul said:
I think Musk is running scared - he doth protest too much, methinks
seems to be trying to enter the propaganda forum to counteract the negative AutoPilot press (fake news?) and get in with the truth, it's like the current anti Russia meme.

The 40% safety improvement he attributes to AP seems controversial
https://www.bloomberg.com/news/arti...tic-in-autopilot-defense-comes-under-scrutiny
it is too commercially sensitive to reveal too much about the systems capabilities, unless there is some legislated disclosure which all vendors have to adhere to.
I would like to see AP1 versus AP2 statistics break out - is his commercial decision to drop Mobileeye compromising safety .. but is he going to reveal that ?


....
Click to expand...

This note from Ars discusses much of the same thing as Bloomberg but the comments are of interest if you have not seen them.

https://arstechnica.com/cars/2018/05/sorry-elon-musk-theres-no-clear-evidence-autopilot-saves-lives/
 
But I thought the whole point was it would be able to avoid these kind of collisions? :p

It reads like another driver took evasive action to avoid a crash, but the AV didn't. So it had a crash. Even though technically it wasn't it's fault, it kinda was.
 
Last edited:
I read it like the other vehicle took evasive action to avoid a 3rd car and crashed into the AV. Sounds like the AV did nothing wrong and was just a passenger in the incident. There is not a lot an AV can do if another car decides to drive into it. As per human driver training they will be programmed to stop, not swerve as that can have all sorts of unintended consequences.
 
Funny enough now that you mention it an AV Should be able to avoid such an accident assuming that there was nothing to its side. The advantage of AVs is that they react faster than humans and have up to date information on their surroundings. Being able to do a "rapid" lane change should be possible considering I've watched videos of humans doing such things.

Unless this AV isn't capable of high speed handling.
 
Chuk_Chuk said:
Funny enough now that you mention it an AV Should be able to avoid such an accident assuming that there was nothing to its side. The advantage of AVs is that they react faster than humans and have up to date information on their surroundings. Being able to do a "rapid" lane change should be possible considering I've watched videos of humans doing such things.

Unless this AV isn't capable of high speed handling.
Click to expand...

I believe you are assuming the AV was in AV mode at the time of the accident. This assumption is wrong. According to a statement released by Waymo, the AV was actually being driven in manual mode at the moment of the accident by the safety driver who was behind the wheel.

Check out the video here and you determine if high speed handling, even if it were in autonomous mode, could have made a difference:

https://www.usatoday.com/story/tech...olved-crash-arizona-driver-injured/582446002/
 
Last edited:
sesevans said:
I believe you are assuming the AV was in AV mode at the time of the accident. This assumption is wrong. According to a statement released by Waymo, the AV was actually being driven in manual mode at the moment of the accident by the safety driver who was behind the wheel.

Check out the video here and you determine if high speed handling, even if it were in autonomous mode, could have made a difference:

https://www.usatoday.com/story/tech...olved-crash-arizona-driver-injured/582446002/
Click to expand...
Didn't realise it was a ahead on. Also that silver driver is blind. Assuming the US has the same traffic light rules as us. They ran a red light and didn't even see the very obvious car to the last minute. Going to bet they were on their phone
 
Running lights seems to be a thing in the US. I saw the aftermath of least 3 crashes because of it when I was in Vagas. Over here it's quite rare to see it (well, except for cyclists lol).
 
You must log in or register to reply here.
Back
Top Bottom