Driverless race cars

[Baatfam]

[Cipher]

not sure about that, certainly our generation would lose interest but there's an entire new generation that watches video game competitions and robotic contests without actually taking part. Hard to predict where this will be in 50 yrs time...electric driver-less F1???

We have a huge number opf people who are mindlessly enthralled playing slots. ultimately all things are just a set of variables.
Sport lovers spouting out stats truly are the worst. They create in their mind th illusion that the whole of the sport they love and its outcvomes are any more meaningful than shooting dice. The fact they commit historical outcomes of crap shoots is sadder still.

[wyly]

OMG! I can't help but think, put that in a motorcycle! lol Or, in a Miata. hehe

1000hp?

https://www.webbikeworld.com/criminal-customs-turbo-help-us/?utm_source=Motorcycle+Newsletter&utm_campaign=5b2cf83ab0-EMAIL_CAMPAIGN_2022_01_04_07_46_COPY_01&utm_medium=email&utm_term=0_b175d17323-5b2cf83ab0-131723573&mc_cid=5b2cf83ab0&mc_eid=7c3f3056ea

[earlfor]

1000hp?
https://www.webbikeworld.com/crimina...eid=7c3f3056ea

I bet that thing is going to be a giggle. heh eh

[bwringer]

I think there will be a place for remote-controlled or even AI racing. Battle Bots was quite popular and entertaining; these robots were remote controlled, but it's not hard to imagine this taking the next step into autonomy, and of course we'll have some "human vs. machine" races.

Race tracks and the sky are fairly predictable environments. As far as civilian self-driving, there are some very large ornery elephants in the room no one seems to be talking about: weather, hardware damage, and human intervention and skill. No one has demonstrated a self-driving system capable of dealing with a routine midwest snowstorm at night, the sort of thing a reasonably skilled human can handle in reasonable safety.

And self-driving also depends on a network of sensors and cameras that are all perfectly clean and functioning perfectly. So if a bug or rock hits a camera, or it it's wet and the view is distorted, does the car become a nicely upholstered paperweight? Lastly, self-driving systems depend on a gossamer-thin legal and functional fiction that a human is paying attention and ready to skillfully take control if there's a problem. Neither of these have proven to be true in practice. Human attention wanders in minutes, and even if the biped is paying attention, it can be extremely difficult to tell when the AI is hurtling toward disaster and it's time to take control. You also have an enormous paradox looming; as self-driving systems get better and better, humans will drive less and less and skills will inevitably atrophy; we'll have humans without the ability to recognize or intervene when there's a dangerous situation, and they won't have the skills needed to continue in bad weather, or if the sensors or AI have a problem.

As far as military uses... yep, getting humans out of combat is an explicit long-term military goal. Remotely controlled surveillance and armed aircraft have been a thing for quite some time; the computer on board handles the flying, and a human at the other end somewhere makes the decisions.

AFAIK, no one rational is yet comfortable giving a computer the ability to decide to kill. But it's not much of a leap to imagine more and more strategic abstraction that blurs the line; "Follow Mr. X using facial recognition, then if you get a clear shot where he's 10 feet from anyone else and there are no children around, put a slug through his left incisor." A human made the decision, but did not pull the trigger. At what stage is there a difference that matters? Hard to say.

They are seeing some very odd and serious psychological effects on the remote pilots. Much of this info is kept secret, but there have been some good articles. When you spend the day tracking and killing targets in Iraq with zero physical danger, then go home to your family in the US, it's far more difficult to cope than anticipated.

At the same time, the military makes heavy use of some very elderly analog aircraft, (the B-52 is expected to serve into the 2050s) and so human pilots will be the most sensible, cheapest option for many, many years to come for many bombers, cargo aircraft, and similar. You'd need a comprehensive retrofit, or a fairly anthropomorphic robot pilot (who will be named "Otto"...). And even then, it's still impossible to recreate human creativity and ingenuity if and when something goes wrong.

Simple things like speech recognition and speech generation for communicating with ATC and other aircraft are still nowhere near the point they'll need to be for full autonomy.

And consider times when human judgement has found a solution no machine could have considered. For example, Sully landing that A320 in the Hudson, or the case of the 737 with dual engine failures that landed on a grass levee near New Orleans. Yes, really: https://en.wikipedia.org/wiki/TACA_Flight_110 (And in this case, the airplane was repaired, flown out, and eventually returned to service.)

Of course, there are arguably more times when human judgement has utterly failed. For example, idiot pilots on a repositioning flight who decided to see what she'll do and cratered a CRJ-200. A computer would have stuck to the flight plan.

https://en.wikipedia.org/wiki/Pinnacle_Airlines_Flight_3701#Investigation

[Buffalo Bill]

I say it would depend upon the mission--speaking generally for combat aircraft, not just fighter planes. I think that for the foreseeable future most military aircraft will with extensive electronic assistance continue to be flown by humans; those humans just won't be on board.
...

Don, are drones landing autonomously now?
Are commercial airliners landing autonomously now? I thought it was common since 2000?

[wyly]

Don, are drones landing autonomously now?
Are commercial airliners landing autonomously now? I thought it was common since 2000?

I wondered the same thing so I did a search on that and I found this.

Can a plane land on autopilot in case of emergency?
– Captain4430, Melbourne, Florida
Many autopilot systems are capable of landing the airplane. This is usually the case in very low-visibility conditions. However, it takes a significant amount of training and knowledge to properly program the flight management computer, properly switch the autopilots into the proper modes and then monitor their performance.
Recently, some small airplanes have implemented an emergency system that will automatically land the aircraft on a runway, after advising air traffic control that there is an emergency, evaluating the weather and navigating to the runway. It is very sophisticated using a lot of artificial intelligence. It is highly likely that this technology will expand.

[bwringer]

Automatic landing has been available in airliners since the 1960s. How automatic is a matter of interpretation; humans need to set everything up perfectly and communicate and coordinate with ATC, and the system manages all the controls.


However, Garmin won the Collier trophy in 2020 for its automatic landing system, now fully certified and available in some smaller aircraft.

https://discover.garmin.com/en-US/autonomi/

Basically, if it detects an incapacitated pilot, or the passengers push a button, it chooses a suitable runway, navigates around weather and other hazards, and lands. The system also handles communicating with ATC and emergency services, communicating with passengers, etc. all the way to a complete stop on the runway. It's really science fiction stuff.


Also, bear in mind that you can buy a drone for a few hundred bucks that can do some extremely advanced maneuvering on its own. For example, lots of adventure riders carry a drone and set it up to follow them automatically as they ride through challenging terrain. The software detects and avoids rocks and trees, while tracking and following the subject at whatever angles needed. Flight control and computer vision software is incredibly advanced.

[rphillips]

Ya'll have got off point talking simple things like flying & landing an airplane. This thread is about driving race cars, really complicated stuff like quick decisions & reflex & anticipation stuff somebody else may or may not do. Shoot flying an airplane ain't nuthin.

[dpep]

Ya'll have got off point talking simple things like flying & landing an airplane. This thread is about driving race cars, really complicated stuff like quick decisions & reflex & anticipation stuff somebody else may or may not do. Shoot flying an airplane ain't nuthin.

Until something goes wrong.

Actually the flying isn't all that bad. It's the takeoffs and landings that can be a b!tch.
...

[RichDesmond]

Until something goes wrong.

Actually the flying isn't all that bad. It's the takeoffs and landings that can be a b!tch.
...

Just avoid trying to break the low altitude record and all is well.

[bwringer]

Ya'll have got off point talking simple things like flying & landing an airplane. This thread is about driving race cars, really complicated stuff like quick decisions & reflex & anticipation stuff somebody else may or may not do. Shoot flying an airplane ain't nuthin.

I think we may be at or very close to the point where an autonomous race car, all else being equal, could better the best available human driver's lap time on a closed road racing course.

I think we're much further from the point where autonomous race cars will reliably win races against human-piloted race cars. Racing is a game of high-speed 3D chess, with all sorts of chaotic and unpredictable things happening, high-level strategic planning, psychology, etc.

And of course the middle step is a race between various autonomous cars; that will actually be quite interesting. The DARPA off-road races were pretty interesting, but the last was several years ago, and the cars were very slow.

[wyly]

I think we may be at or very close to the point where an autonomous race car, all else being equal, could better the best available human driver's lap time on a closed road racing course.

I think we're much further from the point where autonomous race cars will reliably win races against human-piloted race cars. Racing is a game of high-speed 3D chess, with all sorts of chaotic and unpredictable things happening, high-level strategic planning, psychology, etc.

And of course the middle step is a race between various autonomous cars; that will actually be quite interesting. The DARPA off-road races were pretty interesting, but the last was several years ago, and the cars were very slow.

I know F1 already has rules that limited computer aids, they were taking away from driver input, computers don't make mistakes, they brake, shift and steer better than humans.
Computers can react quicker than any human and are immune to psychological tactics they have no fear and can't be intimidated I learned that lesson years ago playing chess with computers. After learning what they could and could not do I manged to defeat those early programs but today's computers defeat even the Grand Masters, calculating, cold blooded, fearless, ruthless, killers.

[Grimly]

the Grand Masters, calculating, cold blooded, fearless, ruthless, killers.

I've known a chess player or two like that.

[wyly]

I've known a chess player or two like that.

lol me too, my best defence was equivalent to holding back a tsunami with a fishing net.