How can I track the location of an object relative to a point?

[the question is at the end, but I want to clarify, first :]

I am trying to interface the control scheme for a helicopter into a 3D representation on a PC.  I was just wondering the best method for tracking the real-world position of the helicopter.

I've considered remembering the starting position and tracking the movements/commands, but I don't know how accurate this would be as there are several influences and variables to consider.

Another idea - probably the easiest - is to put a GPS locator on the helicopter and transmit the coordinates to the computer.  Unfortunately, for the precision and quantity I am wanting, it's a bit pricey.

The other idea is to use beacons and triangulate the position based on distances from the beacons,  but I am not sure how to measure the distance to the beacons.  Ideally, it will be within inches, but I could work with less precise measurements.

How can I reliably/semi-accurately judge the distance from the beacon?  Any other ideas that might work better?

sort by: active | newest | oldest
gmoon6 years ago
Use audio (ultrasound) beacons. Time the delay between transmitting the beacon pulse and it's reception--there's your distance. Then triangulate.
kidmosey (author)  gmoon6 years ago
This seems like the easiest/cheapest to implement. I'll have to play around to see what kind of range and accuracy I can get, though, but that's half the fun. Thanks! :D
frollard6 years ago
The latter 2 options are the best for accuracy. Triangulation with either signal strength or timing will be inaccurate with a homebrew setup -- you can run cameras, and with at least 3, usually more, you can calculate using machine vision the exact location of the chopper. Easiest by far would be GPS, which can be had in the 5-10hz update range for less than 150 dollars.
kidmosey (author)  frollard6 years ago
Thanks for the reply. I think I'll have to go with GPS. Since I'm wanting several of these, I need to keep the prices a bit lower than $150 each, though (<$10 is my budget). Do you have any idea what kind of precision I could get from harvesting a module from a worn garmin or tom tom? Since they are intended for vehicles, my guess is they are not exactly pinpoint; but I can't find anywhere that gives a range.
kidmosey (author)  kidmosey6 years ago
found a page that states <10m for the tom tom precision... Another problem with GPS is elevation, so I guess I'll check out my options with cameras.
Standard GPS isn't precise enough - without extreme methods, its only good to about 30 feet. In the scale of what you are doing, that's horrendous.

You can get GPS modules which are considerably less than 150 USD BTW.

State of the art would be either differential GPS, which is good to inches, but expensive or survey-grade which is incredibly expensive, but good to a millimetre or so

Me, I'd stick with cameras

As per my post, In my experience with a cheap garmin etrex handheld, yes the accuracy is low -- but he's interested in differential data -- if I move the unit just a few cm, it registers precisely how far I've moved. It's not as if the reading jumps about every second. If I place it on a stationary platform, it reads absolutely no motion, even when at 100+ meter accuracy.
That's not my experience over medium distances, say 20 or 30 yards. It would be interesting to see how yours behaves if you remove it say 30 yards, and return it to the same place. Steve
A handheld can get ~3m accuracy. One must understand the difference between accuracy, precision, and resolution. Accuracy is indeed how close the reading is to what the actual position is. Precision is how repeatable a given test is, and resolution is obvious, the quality quantified. While a handheld has 3m accuracy, it's precision is quite a lot better. If you move the handheld a few cm, it will register a few cm (quite precisely), but it is still uncertain within that 3m radius. Handheld units will give you data usually on a 1 hz update. That should be enough to work with -- you can get units that do more number crunching and spit out data more frequently, often extrapolated from the previous datum.