precise location in space

[CeaSaR]

To your question Kurt,

The pros have been using RF for their telemetry for years, but they have the big bucks to do it - several heavy duty computers running the numbers that stream in on the receivers, etc.

GPS is too slow for racing (unless you are out in the middle of a desert or the ocean where you need it to stay on course!) on an oval or closed course. I'm talking about your off the Walmart shelf type of GPS here.

Visual tracking is a possibility, but would have to be tailored to each track and is not a real time solution - lots of post-processing.

If you want realtime positioning, RF is the way to go. I seem to remember years ago a system that was developed for R/C owners to follow their cars over a course. Could have been in the '80's?

As for the Telemetry system, isn't the beacon in the vehicle while the receivers are at the corners of the track? And to make it truly portable, the receivers would have to send the information to a computer over 4 separate wireless channels to be recorded/proccessed. 4 is the number used for high accuracy, but it can be done with just 2 receivers placed ~ 10' apart on a horizontal pole. I know it can be done, but at what cost and would you be able to get it to work reliably enough while your son is still karting. In addition, wouldn't you need approval from the track owners/sanctioning body for use of the equipment to be placed around the track?

Just a few thoughts culled from the many flying through my head...

CeaSaR

[cato]

Kurt -
Do the tracks have side walls that you could bounce ultra sound off? That would give you a side to side measurement...of course you would have to make provision for the other vehicles blocking your measurement.... Could you add an encoder to the wheels to measure linear distance...you could, i suppose also measure steering position combined with your speed/distance measurement to dead recon the position of the car relative to say the start line...alternatively, or additionally, an electronic compass with enough resolution/accuracy might be found to give you heading info to combine with your speed/distance measurement...

Engineer-
I'm under the impression that the error goes away with time. That is, as you average more and more measurements your accuracy increases. I have seen professional surveyors using GPS mounted to their tripods.

[Engineer1138]

OK, so maybe you can answer a question for me

The survey shows an [x] and the key reads "wood post/metal spike at 11 foot offset" What the heck does that mean? I looked 11 feet from the corner of the house where the dotted line was, but no post/spike. I spent a fair amount of time online this weekend reading about surveying, but that was never mentioned.

I live on a 10-acre plot and the corners I'm most interested in should be about 1,000' from the house, hence my thought that GPS was the most feasible solution.

Originally posted by CeaSaR.:
Engineer - all you need is a long tape measure, a scientific calculator and a metal detector to find your property corners. Since offsets are supposed to be at 90 degrees, use the Pythagorean Theorem. I ought to know, I've been doing that for a living for 23 years now. It is surprisingly effective.

However, if you have a really large parcel then you could use a consumer grade gps handheld to get you within 6-10 feet +-, then use your metal detector to find your corners - assuming you are oriented correctly.

[Newz2000]

Originally posted by CeaSaR.:
If you want realtime positioning, RF is the way to go. I seem to remember years ago a system that was developed for R/C owners to follow their cars over a course. Could have been in the '80's?

The local hobby store has an r/c car race track and they can tell which cars are on the track and how many laps they've done. It's all run via a laptop. It looks fascinating, but I haven't had the time to figure out how it works yet.

I wouldn't loose hope, because it seems there's some way to do it.

BTW, I read an article a few years ago that explained how race cars now had cameras on them that looked at the ground. They could use image data of the ground to see how they were turning and performing. It was based on the grain of the pavement or something similar.

[haklesup]

Visual tracking using astereo vision camera should work and have the byproduct of a video record.

It might be pricy per the link but there must be a hobbiest approach using a couple of USB webcams and some public domain software. There were lots of non commercial links when I googled "Stereo Vision"

[kheston]

Hmm, stereo vision. I'll take a look.

[Dimbulb]

If you have to monitor him that close he should'nt be driving. He is not a dog.

[dyarker]

I think the objective is feedback on driving skill and finding the fastest path around the track, not monitoring. On Formula 1 cars every turn of the engine, and every bump in the track, is transmitted back to computers behind the pits!

[haklesup]

Another 2 cents worth:

If using a vision based system, you can tinker with the contrast so that the only thing the camera sees is the target. I was thinking that you could put a dark red filter on the camera and a bright red LED on the car then turn up/down the white balance/contrast until almost nothing but the LED beacon is detected by the camera. Now the software can much more easily pick out the target among the other visual clutter (as fully saturated pixels).

You may even have luck with a IR using a remote control LED and filter, most B&W cameras are sensitive to near IR. It should give a better SNR and reduce the video processing requirement.

[jwax]

I'd look into the ultrasonic approach-
Rangefinders are available and several of them around the track could feasibly be "locked" on a beacon that is on the kart of interest. Rate of change of distance with time is velocity, and from when distance from a base to kart is minimum, the kart is passing that base.

[kheston]

Dimbulb, races are won with improvements that can be measured in fractions of a second. Seeking another competitive advantage makes me just like every other racer looking for a competitive edge. Not to mention it's fun to problem solve just for the sake of doing so.

CeaSaR, you gotten my hopes up again re: the RF option...

My original idea was that I'd use multiple beacons and a data logger on the kart, but I think your idea is better, multiple stationary receivers would be more stable.

I don't think the video option would be as flexible. The main track we frequent has no walls bordering it. It's pretty big, too. That means multiple cameras. I suppose I could just float above the track in my helicopter, if I had one, with one really good wide-angle camera. But then I'd probably be able to afford some military-grade RF positioning components, too.

I think I can cover the data processing part, it's what I do. I wouldn't necessarily have to process the data while we were at the track, that could be done at home. So, perhaps a bunch of receivers and a dithering algorithm might get me the accuracy I'm looking for? Where the best place to start looking for a set of beacon and receiver circuits?

I've never played with a range-finder. Are prices for decent ones in the hundreds and not thousands?

<small>[ February 01, 2006, 07:14 PM: Message edited by: Kurt Heston ]</small>

[rshayes]

A beacon on the cart with multiple receivers could be used to determine the beacon position by measuring the relative phase of the received carrier at each location. A 1 mhz signal has a wavelength of 300 meters, which might be enough range. Higher frequencies would allow higher resolution, but would also require keeping track of phase ambiguities. Changes in the overall path caused by other moving objects might limit the accuracy which could be obtained.

Absolute position information would not be much use unless the borders of the track are measured with equivalent accuracy. This would either require a good map of the track or extensive calibration.

Cameras on the cart could provide the information you are looking for if the video images can be processed to locate the track edges. A pair of accellerometers on the cart might also give useful information, for example, if the turns are being taken with the maximum lateral accelleration. This might even be of use in real time, if the optimum combination of velocity and turn radius can be calculated. A double integration of the accellerometer outputs would also give you the actual position of the cart.

Map following systems have also been built using the distance a vehicle travels combined with direction information from a compass. Slipping wheels would limit the accuracy of this method. A velocity sensor based on an ultrasonic doppler shift method might be more accurate. For example, this web page (http://cigr-ejournal.tamu.edu/submissions/ volume3/PM%2001%20007.pdf) is a paper on the subject.

<small>[ February 02, 2006, 12:51 AM: Message edited by: stephen ]</small>

[jwzumwalt]

The technology that use to be used in aviation is "phase angle" change. Essentialy, you broadcast a long wavelength signal from multiple stations synchronized and sum or subtract the recived signal to see how far of the sine phase they are - this becomes a fraction of the distance ratio.

For example, ground proximity radar uses a simular idea. FYI the gentlman that suggested using the aviation package made a well meaning simple error - using aviation radio equipment in anything but a certified aircraft installation will get you some time in Federal Prison

Another technique used by aviation that could be used at some other frequancy is the technology in use for VOR sites. A descrete carrior burst is broadcast once per second (consider this the (time mark). Then a second single is directionaly sweeped through 360deg. For example, if you recive the sweep signal and burst at the same time you are north (0 deg) from the station. if the sweep is 1/4 second after the burst, you know the sweep signal is being pointed 90deg from north, reciveing the directional signal 3/4 of a second after the burst means you are 270 deg from north, etc. Using this technique only requires two transmitters for triangulation because you are being given relative direction information, not distance.

[Gorgon]

Hi Kurt,
Your comment about the helicopter made me think of another way to get airborne pictures. What about one or more small Helium filled blimps with one (or more) camera onboard. You could then put out markers on the ground for reference and whatever automatic control of the camera(s) you want. With a downlink and some picture analyzing software you could solve your problem? It should be possible to make stereoscopic images too, for 3D analyzing.

And, no problem with moving to a different track

TOK

[kheston]

Update!

A buddy of mine is a heavy-equipment operator and uses a GPS-based system that directs him as to how he needs to set his dozer's blade (angle, depth, etc.) such that he can move earth without the need for surveyor's markers and such. Apparently, it's accurate to the fraction of an inch in 3-dimensions.

I'll do some research and see if I can figure out what extra mojo they're using to divine that sort of precision into GPS. More to come...