precise location in space

[kheston]

My son is a karter. I'd like to keep a better account of where he is on the track at any moment in time such that we can refine his line and do other coaching based on the data we log.

There are kits out there that measure how he's using the kart's controls, but his position on the track isn't quite as measurable.

GPS is far too inaccurate for this application. A couple of inches of error might be okay, but GPS is usually off by the tens of feet.

Where would I start looking for components with which I could use radio beacons (three or four of them) and a receiver to triangulate where the kart is at any given time? Would I ever get the accuracy I'm looking for?

I'm a programmer with a grasp on mathematics and could do the crunching of the raw data, but I'm having trouble figuring out how to capture it. Any suggestions?

[Chris Smith]

Try one or more cameras.

All the other sports use them as a teaching method.

One corner at a time, or multiple ones if you can afford many.

Plug them in and go.

[kheston]

Cameras leave too much to one's interpretation. I'm thinking I can be more precise.

Let me be more simplistic. What is the easiest (yet most accurate) way to use radio or another wavelength to measure distance?

[kheston]

Some Googling yielded this:

http://www.eng-tips.com/viewthread.cfm? ... 043&page=1

The twin-GPS idea is an interesting one. I've got two Garmin eTrex that might work. Hmm.

[Chris Smith]

Cameras and OR lasers can triangulate it best.

The "interpretation" part can be completely removed by placing back drop sheets for a measurement grid that allow you to visually measure and triangulate exactly where it is, to the inch or less.

Patterns striped like a Zebra [or checker board] with numbers on each stripe allow the camera to be blocked by the object, yet count measurements from left to right or vice versa like the Height sign at a 711 store. And you only need cover one area in question at a time.

Multiple angles and cameras [2 or more] corroborate the findings more accurately, and the props can be moved from one place to another to examine each corner one at a time.

Card board cut outs work well, and wont add in any risk of crashing into them.

OR,... you can also do it with lasers, but that’s a little harder. You need a beam splitter and multiple sensors that get blocked, and these are at fixed points so when one is blocked, its absolute fix, is given to a exact spot and it is 100% accurate down to the diameter of the beam.

[jimandy]

I'm with Chris - lasers or cameras. RF triangulation is NOT the way to go. Way too complicated and expensive unless you can buy an FAA type aircraft Omni system. If you could get the cooperation of the Kart course people and other Karters you might come up with something of more universal interest and value (and share in the expense). Perhaps an underground system of grid sensors with each cart telemetering back their position.

<small>[ January 29, 2006, 04:13 PM: Message edited by: jimandy ]</small>

[kheston]

My system needs to be pretty portable. I plan to use it at more than one track.

So for RF, I'll need to spend a mint on electronics, huh? Schucks.

I had envisioned coming to the track, strategically setting up a few tripods with beacons on them, and commencing with laps. Oh well, when my grandkids show me a working system like this in 50 years, I'll tell them I thought it up a long time ago. Of course, they'll have fusion powered karts by then, too.

[Chris Smith]

Visual triangulation is the key to simplicity, portability, and cost.

[Colinr]

Gps accuracy can be inproved using differential GPS

it requires haveing a fixed GPS base station that reads in the GPS and compaires it to its actual position the error is then transmitted (via rf link ) to the mobile GPS receiver on the Kart the kart gets its position from it's own GPS receiver and then uses the error information from the base station to correct the result.

Note if you are looking for post recording data analisis you may be able to record both sets of GPS data seperateally and then combine them as part of the analysis.

Colin

[Engineer1138]

On a related topic, can anyone tell me how frequently GPS error changes? I'd like to try a "really poor man's DGPS" by measuring the error at a known location, then running off to the unknown locations, getting a GPS reading, and applying the error correction.
Or am I way off base here? Does the error change wildly from second to second?

The application is finding the perimeter of my property. I just bought a new house and while we have an official survey (2 years old), I can't find any of the survey monuments (all that white stuff on the ground doesn't help either so I want to try using a corner of the house as a reference and finding the corners of the property by offsets from the distances shown on the drawing.

[Chris Smith]

The military doesn’t want you to use their equipment that accurately so they built in errors to stop you from making your own G.P.S. guided smart bombs.

However for the outdoors person and the road map reader, this error doesn’t come into play badly.

But down to the inch Im sure will always be reserved for the military toys and not the public, while the error is random to keep it that way.

[Engineer1138]

I believe Selective Availability was turned off a few years ago. The errors I'm talking about are the computational errors in the receiver due to drift, poor reception, etc.

[ljbeng]

RTK GPS is the most accurate, <1cm but is the most expensive and can be used by anyone. You have 1 base and 1 rover and X,Y,Z is all very precise. SA is off and doesnt' affect RTK anyway.

1138, I have experimented with what you are asking and no, it does not yield you any more accuracy.

A WAAS based GPS will get you a lot more accuracy than a GPS alone but not near what RTK can.

Omnistar corrections from their own satellites could get you < 3 ft accuracy but that requires a subscription and equipment to receive the corrections.

Are you near any US CoastGuard beacons? These corrections are free with required equipment but I don't know the accuracy you can get there.

<small>[ January 30, 2006, 11:50 AM: Message edited by: ljbeng ]</small>

[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.

CeaSaR

BTW, don't forget to obtain a map of your subdivision or at least the surrounding neighbors deeds - all available for copy cost at your respective Recorder of Deeds office. they offer a heck of alot of clues as to the whereabouts of your mischievious boundaries.

Good luck!

[Chris Smith]

I believe Kurt is looking for accuracy down to the inch give or take one and perhaps three?

As a racer in my teen age years well before the dinosaur roamed the earth, the angle into a corner, the landing spot, and the return are all part of a parabola that is best made down to the inch for the maximum speed in and out of the corner to reduce your momentum and drag against gravity and momentum.

Several inches, as in 12 to 36, doesn’t teach you anything because ALL of the competitions inaccuracy is well within these limits.

The only way for a poor mans parabola to be charted is through the use of optics,.... be it the camera or through the use of lasers, also optics.