Self-driving vehicles.

Advertise with us
shop deals on amazon
SilicdyneTRF
The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Most people are familiar with conventional cruise control, and most people can imagine what level 5 autonomy (real, full self driving) would be like (doesn't exist yet). So here's a good survey of features in between: more advanced than regular cruise control, but still a long way from reliable hands-free driving:



Tesla's "Full-Self-Driving" software adds to the above for $12k or so, allowing for attentive, hands-free driving.

It's all about pushing safety to the extreme.
 
Another company working on self-driving. Regardless of how quickly or closely they reach fully-autonomous driving, first uses are in smaller, tightly controlled areas. An airport looks like a great example. Vehicles can easily be monitored or controlled from the tower.

https://navya.tech/en/
airport.jpg
 
cwbullet said:
I don’t know that I will ever trust a vehicle to drive itself. There are so many variable in driving safely.
Click to expand...
It's a very gradual process of adding and exhaustively testing many features over many years. Cruise control is the most familiar example. No one really disagreed with idea of automatic braking. And I guess it went on from there.

Some underlying ideas are that a computer can react quicker than a human, and has a 360 deg visual range. One measure of success is the number of accidents. Hard to disagree when it's lower that human driven cars, or when it prevents accidents from happening.

In any case, I see no problem with testing anything at low speed on closed circuits.
 
Funkworks said:
It's a very gradual process of adding and exhaustively testing many features over many years. Cruise control is the most familiar example. No one really disagreed with idea of automatic braking. And I guess it went on from there.

Some underlying ideas are that a computer can react quicker than a human, and has a 360 deg visual range. One measure of success is the number of accidents. Hard to disagree when it's lower that human driven cars, or when it prevents accidents from happening.

In any case, I see no problem with testing anything at low speed on closed circuits.
Click to expand...
My car auto brakes. It works well if someone stops suddenly.
 
cwbullet said:
I don’t know that I will ever trust a vehicle to drive itself. There are so many variable in driving safely.
Click to expand...
It is likely to be a generational thing. At some point there will be kids who grow up with self-driving cars and it will seem normal. Those who have spent a lifetime driving will have a more difficult time adjusting... but eventually we'll all be dead. :)
 
neil_w said:
It is likely to be a generational thing. At some point there will be kids who grow up with self-driving cars and it will seem normal. Those who have spent a lifetime driving will have a more difficult time adjusting... but eventually we'll all be dead. :)
Click to expand...
True but depressing.
 
neil_w said:
It is likely to be a generational thing. At some point there will be kids who grow up with self-driving cars and it will seem normal. Those who have spent a lifetime driving will have a more difficult time adjusting... but eventually we'll all be dead. :)
Click to expand...
Yes, but will we be dead in a fiery crash? :D😬

I thought this article was interesting--drivers overestimate how much self-driving their existing cars can do.

https://arstechnica.com/cars/2022/1...restimate-hands-free-driver-tech-study-shows/
 
cwbullet said:
True but depressing.
Click to expand...
Oh hey woah um ... No! 😁

If they reach full-driving (and it it might take a generation just to reach level 5 autonomy), it will be a new option, not an overnight replacement. People will gradually try it, and when it's good enough, people will voluntarily switch over because they will find it to be a better option. At least that would be the goal of the engineers making it. That's how technology usually stays: those who try it don't want to go back. Like smartphones over rotary phones, rocking chairs over logs, sneakers over sandals, etc.

boatgeek said:
I thought this article was interesting--drivers overestimate how much self-driving their existing cars can do.
Click to expand...

Over-estimating can result from some people overselling it. You have some of them everywhere but it usually isn't those who do the actual code and engineering. My view is that as long as people have ideas as to what to try to solve the issues, the ideas are worth trying. But it's hard to say how much time it will take. Seeing progress in AI and self-driving will be interesting even if it takes decades to reach the ideals like what they have in sci-fi movies.

In any case, whatever happens in the public eye, there's always a core gang of developpers working on it behind the scenes. AI has made big leaps over the last decade and there's a "million" things to try and see how far it can go.
 
Last edited:
cwbullet said:
I don’t know that I will ever trust a vehicle to drive itself.
Click to expand...

People used to say the same thing about elevators when operators were being phased out. They were sure there'd be a lot of people cut in half as the machine failed as people were stepping on or off. Without a human there to make decisions a machine couldn't be trusted.

But as was pointed out, once the safety record is looked at, it's hard to argue with the collisions per mile of humans vs computers.
 
I'm interested in how they'll develop the ability to auto-drive without the benefit of system reliant ground sensors or centralized monitoring/control and how that will alter individual movement and societal impact.

 
Funkworks said:
Oh hey woah um ... No! 😁

If they reach full-driving (and it it might take a generation just to reaching level 5 autonomy), it will be a new option, not an overnight replacement. People will gradually try it, and when it's good enough, people will voluntarily switch over because they will find it to be a better option. At least that would be the goal of the engineers making it. That's how technology usually stays: those who try it don't want to go back. Like smartphones over rotary phones, rocking chairs over logs, sneakers over sandals, etc.



Over-estimating can result from some people overselling it. You have some of them everywhere but it usually isn't those who do the actual code and engineering. My view is that as long as people have ideas as to what to try to solve the issues, the ideas are worth trying. But it's hard to say how much time it will take. Seeing progress in AI and self-driving will be interesting even if it takes decades to reach the ideals like what they have in sci-fi movies.

In any case, whatever happens in the public eye, there's always a core gang of developpers working on it behind the scenes. AI has made big leaps over the last decade and there's a "million" things to try and see how far it can go.
Click to expand...
Read again and you will see what I meant.
 
cwbullet said:
Read again and you will see what I meant.
Click to expand...

I think you mean read this:

neil_w said:
... Those who have spent a lifetime driving will have a more difficult time adjusting... but eventually we'll all be dead. :)
Click to expand...
As for the first part, I often agree with neil_w but not here. The holy grail of self-driving is that people don't even need a license, like sitting in a passenger plane (some of which can land by themselves). Anyone from any age and any health level would be able to hop in and be driven anywhere. A car so safe and easy to ride I could set a Fozzy Bear muppet in the driver's seat, tap a spot on the map, and Fozzy could spend his time telling dumb jokes online instead of looking where's he's going 🐻. I'm not sure how the last part relates👻.

Banzai88 said:
I'm interested in how they'll develop the ability to auto-drive without the benefit of system reliant ground sensors or centralized monitoring/control and how that will alter individual movement and societal impact.
Click to expand...

I'm not aware of any car of any kind using ground-specific sensors (self-driving or not). If I take Tesla as an example, what they're going for now is all-around optical cameras as sensors. Cameras see the road ahead like a human does, including pot holes, as a human would, and analyzed with AI available (AI can distinguish one breed of dog from another in an image). Video footage is taken from the 160,000 participating cars, and sent to a supercomputer that redistributes the knowledge to the same 160,000 participant cars. There is no centralized "control" or "monitoring" per se, but there is centralized AI that collects data to keep improving the 160,000 cars. When a car sees and solves a first-time situation, the solution becomes available to all other cars. Tesla is currently building their own self-driving specific supercomputer with in-house hardware (and software of course) and it will among the most powerful in the world (#1 specifically for self-driving with no one coming close - Intel and Nvidia hardware is not specifically designed for self-driving). So I'm very much looking forward to see how this progresses.

"Tesla is starting to put together a massive, custom-built stack of hardware called Dojo to train its AI on all the video its cars are picking up and beaming back to the company. To get the performance the AI team needs to churn through a 30-petabyte footage vault, Tesla went dense with its hardware."

https://www.cnet.com/roadshow/news/...-we-learned-tesla-bot-dojo-full-self-driving/
 
Last edited:
People need to see it both work for a while and be affordable. I don’t want to pay for a science experiment.

Engineering says it will stay in the lane and not hit the car in front of it.
Marketing tells customers fully autonomous, get in and it takes you there.
Crashes blamed on engineering.

CB34D262-B9BB-489B-8BCD-18444519C7A2.jpeg
 
MClark said:
People need to see it both work for a while and be affordable. I don’t want to pay for a science experiment.
Click to expand...
Yeah, but at this point there is no fully autonomous vehicle. All that is available is levels up to 2 or 3. I think Mercedes' system only works below 30 km/hr, and Tesla's requires a fully attentive driver behind the wheel and the car will wake him up if he looks away, and even disable the feature with repeat "offences". Let's be clear that actual, level 5, full self-driving does not yet exist. The Cruise taxis are restricted to a certain area of San Fransisco during the night and look like Ecto-1 with gimmicks on the roof. Safe progress will take time and seeing how close we can get is part of the fun.

Tesla bot displayed at a Chinese trade show. When complete, it will basically be some level of "Full-self driving" on legs instead of wheels.

 
MClark said:
View attachment 547734
Click to expand...

Hey that's a Marvel character.

https://ideas.lego.com/projects/ed42cf5e-261e-4576-8680-4632b4889999
In case you want a self-driving truck, here's a few links to look into:
https://www.einride.tech/autonomous/vehicles/
And this company will probably look into it if they haven't already.
https://www.tesla.com/semi
👻

🤖🚚🤖🚛🤖🚚🤖🚛🤖🚚🤖🚛🚚🤖🚛🤖🚚🤖🚛🤖🚚🤖

Here's a short one to get an impression of where FSD is at:



Don't anyone get this wrong, it's incredibly hard work to "solve FSD". They don't need naysayers and haters on top of it. It will always be a minority who are brave enough to tackle things that haven't been done before.

:computer:
 
Last edited:
Last edited:
An Army truck driving without any road to follow would be interesting. Most basic would be first vehicle to a location such as a launch site on a playa, minimum obstacles. Across a forest would be a challenge.
They will figure it out.
 
I've been eagerly wainting to see this statistic:

"According to this report, the average Tesla equipped with FSD Beta, driven on predominantly non-highway sections of road, crashes 0.31 times per million miles, a dramatic decrease from the average American, who crashes 1.53 times every million miles."

https://www.teslarati.com/tesla-fsd-safety-statistics/
First clear demonstration I know of that a robot can drive more safely than a human.

🤖🏆
 
Last edited:

Similar threads

Funkworks
  • Locked
Government mandates, forcing and coercion.
2 3
Replies
86
Views
3K
cwbullet
cwbullet
Anonymous
Optimizing Fin Shapes in Rocksim
2
Replies
31
Views
2K
Brandon.L
Brandon.L
stealth6
For the cyclists here (who might actually care about such things)....
Replies
22
Views
472
Kabulrocks
Kabulrocks
5
Korolev Countdown?
Replies
4
Views
722
5thDay
5
StreuB1
Experience with Tapentadol (Nucynta) for pain management?
Replies
1
Views
702
Joshua Smith
Joshua Smith

Latest posts

Back
Top