K.I.T.T. Technicalities: part 2

[Vchat20]

i figured since ever since part 1 there have been quite a few developments in technology that would improve the possibility of a real working kitt, this deserved a part 2 of the debate.

to start off the debate, id like to discuss the possibility of a self driving car that can also navigate. there was a story a while back on TechLive where either honda or nissan had a car that could parallel park itself. take this ability and add highly detailed and highly accurate maps of roads, and this could very well become a reality.

[Knight Industries 3000]

I don't think we'll see a reliable self-driving automobile for many decades, if ever. A robot car would need advanced artificial intelligence just to stay on the road. Many of the self-driving cars that have been developed thus far can only operate under very restrictive conditions: for example, they follow the painted lines on the road. Not all roads have painted lines, and not all traffic lanes on all roads have the same width. Add to that the complexities of interpeting traffic lights, stop signs, recognizing the presence and trajectories of other moving vehicles -- and people and animals -- variable road conditions, weather conditions, lighting, etc., and one begins to wonder how human beings manage to drive!

[Vchat20]

well, the part about detecting nearby vehicles is a very close possibility. if I remember right there was a car that is actually out now in production form that has an adaptive cruise control which utilizes a radar to detect nearby vehicles. this way if you get too close to a car in front of you, it slows down and so on. traffic lights can easily be done by maybe adding an rf transmitter or even better maybe an ir transmitter on the lane which you are stopped at which constantly beams a code which the car then receives and translates to a traffic light color. multiple ones could be placed say 500 feet back in each alane so that even with long lines of cars, each car knows the color. of course this would require massive changes in the road system, maybe someone else has a better design. i was originally thinking of using a close but modified version of either the system that traffic lights use to detect waiting cars or the system that emergency vehicles use to change the lights at will. stop signs are a completely different story but could very well use the same solution as for traffic lights.

[Michael Pajaro]

The technology to build a driverless car may be closer than you think. In March 2004, the defense department is holding a race to see if anyone can build a driverless car that can navigate itself from Los Angeles to Las Vegas in under 10 hours, without any human intervention.

http://www.msnbc.com/news/958593.asp

A few comments however:
This race is being run under very strict monitoring and somewhat controlled conditions, so it will be a VERY long time before you'd see this kind of functionality in a standard production vehicle.

The ability of a computer to drive a car does not necessarily require human-like intelligence. Don't confuse fancy image processing and GPS algorithms with true artificial intelligence.

[Vchat20]

wow. first time ive sen something like that go on. i did a search on that topic just to see if KITT or knight rider was mentioned anywhere. lol. but if thats the case and it does work the way its supposed to, than that would definitely pave the way to a real life working kitt replica. but i still see nothing mentioned about stuff like traffic lights, stop sigs, and exactly how it would navigate itself all the way from LA to Vegas.

[Knight Industries 3000]

The ability of a computer to drive a car does not necessarily require human-like intelligence. Don't confuse fancy image processing and GPS algorithms with true artificial intelligence.

I'm not. There's also a difference between recognizing a stop sign using image processing and determining how long to wait in a stopped state at that sign before it's safe to proceed, determining a safe rate of acceleration given road conditions, etc.

In response to an earlier post, the self-driving car described would seem to require modifications to our road system -- signs, traffic lights, etc. That's a far cry from a KITT-type self-driving car, which presumbly would require no modifications to the existing road system.

To be fair, there's certainly a difference between a computer that possesses sufficient artificial intelligence to drive a car, and a computer that possesses sufficient artificial intelligence to talk, enjoy classical music, and form a friendship with its driver.

[KFCreator]

Actually, I read an article in the San Diego Union Tribune over 10 years ago that was spotlighting a protoype car that was driverless. Interestingly enough, it operated using built-in sensors and was run on a test track that had magnetic "disks" embedded within the pavement that contained all the necessary information such as speed limit, road signs, traffic signal timing, etc. and it was read by a sensor that was underneath the car and picked up this information whenever the car passed over it. The information read by the onboard computer would then display the information on a dashboard LCD screen. They said that in the future it may be possible to use similar technology on all cars, which would eliminate the need for road signs, traffic lights and even billboards (advertisements could be downloaded to these in-ground "disks" as well). Imagine that, a world where you drive down the road and don't have to see advertisements for McDonald's every five feet, don't have to look out for lighted traffic signals and don't have to look for signs all over the road. One can only hope...

[ShadowConvoy]

No, because then we would be bothered with "pop up" ads on the LCD or better yet the windshield!LOL [just kidding]

[neps]

Then we would just need a organization like Mozilla to come out with an alternative program to block the pop up ads.

These ideas of magnetic discs to do things has been mentioned before, but I'm not sure how much I can believe their ability. I remember when BTTF 2 came out, the idea of being able to hover like on the HoverBoards was that magnetic discs are embedded in the side walk, and the magnets in the board create a reverse polarity to force them apart much like putting two of the same sided magnets together. Both ideas sound good in theory, but I just don't know.

[Vchat20]

yu know a very simple resolution to this would be to make a simple modification of the reflectors that are already embedded in the roads. tiny transmitters can be embedded in them and sent to the cars.

[neps]

Sounds simple, but where would their power source come from? Would they then need to tear up the road in order to power them? How does the information get to them, do they need to set up a satellite system in order to get the information there? It's not so simple sadly as one would hope.

[Vchat20]

actually, ive seen something similar to this before. the basic idea was to have powered LED's in them and sensors so that the lights in front of you were white and the ones behind were red. or something like that. they ran off of a battery, but were solar powered also which charged the battery also.

[KFCreator]

it could be more simple than one might think. Consider that there wouldn't need to be that many disks embedded within the roads for some distance unless you were na city, and even then you could probably contain all the information within at least a few blocks needed to transmit to the cars as they pass over each lane. I do like the idea of modifying road reflectors, and it's true that they could be simply solar powered, with and contain an internal battery that would store power for rainy or cloudy days. As for the disks being magnetism-based: I don't know if that would really be the way to go but I'm sure there are other alernatives that might do the job just as well. And a for hw information would get to the disks, maybe some could be pre-programmed with only road information (which usually doesn't require major updating unless the road conditions change or something) and some could be used for just advertisements (which would use satellites and would be run by the companies whose ads are being sent out and displayed). Modifying the roads and maintaining these new disks would obviously require road crews but with most power sources being able to last a remarkably long time these days, it's possible the disks could be run for up to 5 or maybe even 10 years without major work having to be done to them. It really wouldn't be any more busy (possibly it could be even LESS busy) than the road crews you see out today who have to constantly go out to repair signs and fix traffic signals. Anyway, just my optimistic analysis of the possibilities.