I think you may be on track with triangulation using broadcasts from the two devices. I can't be of much help there. You might check into GPS. I believe they can be accurate to 3' last I knew (available to citizens). Maybe they do better now.
You may be able to play games with broadcast power so that when one robot got 10' or so away it would lose the signal and you could light your LED or something when that happens. This would depend on what frequency you decided to broadcast at (may not be feasible given the short distances). I can't help much more with broadcasting signals etc without revisiting some text books so good luck.

The presence of obstacles could not only limit laser, (most) acoustic, IR, and radar sensors, it would also create a false alarm problem - how would the robot discriminate between an obstacle and the other robot? You can avoid these problems with GPS.

Single channel GPS receivers are now available the size of a quarter or even smaller. With GPS selective availablity (S/A) now turned off, their absolute accuracy will be on the order of a few feet, however their relative accuracy should be much less than your 2 foot requirement. You then need a pager-sized or single board serial radio to relay the 1 PPS GPS text message from one robot to the other (or from both to a monitoring station), and a few lines of code to parse GPS messages from both receivers to determine relative distance (and bearing and height, if you want it). Add timestamps and a bit more code to improve accuracy if your robots are travelling more than a foot per second. If the robots are operating indoors, a GPS repeater can be obtained for about $1000. Being a mechanical engineer, you will probably want to avoid writing code; it may be overkill but LabVIEW (see www.ni.com) allows you to do everything graphically, and comes with a few serial comms sample programs that can be adapted to your needs (or you could sub the work to a software type).

This thread is also applicable to a device I'm attempting to produce that will measure distance from a transmitter to a receiver (or multiple transmitter/receiver pairs), but I need it to have higher resolution than 1 foot per second.  Where can I get more info on GPS repeaters and their accuracy and speed?

Thanks,

jwm

Look at technology developed by Bluesoft. It is implemented on Bluetooth and is demonstarted successfully. You can buy evaluation system to test your application. The specific problem you describe will be effectively solved with the 0.5 meter accuracy they provide. You may contact me at: motib@pacbell.net

I am looking for a dependable way of measuring distance on land based drilling rigs. Basically I need a good feedback signal to control a hoisting system on substructures. The distance is 15'. We raise the structure with (6) hydraulic cylinders. I need a rugged and dependable way to do this that can live in the harsh enviorement of the drilling industry. Contact me @ Larry.Desmit@natoil.com

Hi
I too am looking for a microwave distance measuring device but with an accuracy of 1cm. in approx 30 metres. I know it should be possible because I have read articles where this accuracy can be obtained from space! Nevertheless I am a novice and want to measure the distance to a reflector which may move slightly. Being a novice I'm afraid that I can't help you with hardware, sorry.
Regards
Davidsearle

Hi- I am also after electronic distance measurement - to measure distances in the range of 100mm with the sensors 500-600mm away - need accuracy in the range of about 10mm.  Can anyone advise me of source of sensors or of complete kits (I am in Melbourne Australia)such as are used in cameras for auto focussing - Are they accurate enough??

Desmmitlm, You can get hydraulic cylinders with sensors to give feedback to the location of the cylinder. This would tell you the distance. It is embedded in the cylinder so there is no problems with the environment. There are other sensors that would work too. Try Banner Engineering amongst others.

Jmurray
Can U tell me more about your application/need ?

<nbucska@pcperipherals.com>

JJellin,   I wonder if the collision avoidance systems would work for you - I noticed from an internet search that they are available in kit form from the UK.  I don't know the range they work over, but I do know that they are directional so you would need to have several pointing in different directions to cover a whole region.
  Cheers,   Overland.

Hi.

The problem discussed in this thread can be resolved by several ways. It looks like the phase method of distance measurement is the most suitable one in conditions presented. Let’s see on the task of jjellin, and count one robot as a parent that controls his child. Parent radiates RF1 carrier signal, modulated with frequency F1. Child is only a relay. It receives this signal, filters F1, modulates with this the responding carrier RF2, and radiates this RF2 back. RF1 and RF2 are necessary to avoid an interference of transmitted and received signals on the parent side. The first robot receives RF2, filters F1, and estimates the phase difference between radiated and received signals on F1 frequency. This difference will be proportional to the doubled distance between parent and child. Small distances – low needed power radiation, simple transmitters and receivers. Phase measurements can be made just on a base of binary counter. It’s not necessary to use GPS. You can create your own Local PS. Every device on robots will have a size, even less than pager. The shorter wavelength of F1 signal, the more accurate measurements. However, to avoid ambiguous measurements (It’s impossible to measure phase difference that is bigger than 2*pi), the F1 wavelength must be not less than two maximum distances between robots.
Obstacles. If they are transparent for radio waves, like a man for instance, they can give additional undesirable phase shifts. But these errors are able be in required limits. If this is a non transparent obstacle, like a RoboCop, it will ruin the functionality of system. To handle this problem you can use a couple supervisors allocated above robots. In this case every robot will be a child. Two upper parents will measure distances to every child and you will have a triangulated external measurements that will allow you to control all this kindergarten.
Another problem can be if this unambiguous system isn’t able to provide required accuracy, like it’s necessary for JMurray and Desmitlm. In this case you must to build system with ambiguous measurements instead. That means the every distance must be measured utilizing several wavelengths F1, F2, …Fn. The number n depends on the requirements for system. Every wavelength can be much less the maximum measured distance and will give ambiguous phase difference. But it’s possible to process them all together to resolve this ambiguity and to get accuracy in centimetres and even less on long distances. Ambiguity resolution isn’t a problem; it just will complicate parent devices a bit.
In case that Desmitlm have, the problem can be related with presence of many metal constructions in the region of interest. They will give lots of reflections of radio waves. But this problem can be resolved by choosing the positions for parent and child points. Another task – to make system able to survive in the harsh environment is a constructive task and can be resolved as well.
So, everything is in our hands.

Regards,
You may contact me at: dirfind@hotmail.com

RFID is the answer. I do work with company that provides advanced Radio Frequency Identification (RFID). With this product, place a RFID Tag on the object. Using a PCMCIA card in a laptop/PC or handheld device, ping the particular robot and when it reaches the outer zone where they will loose communication. A LED blinks on the TAG during communication, and both visual/audible indication is on the scanning device. The power level can be set in many combinations and can read from virtually 0' or reach up to 300'. For your application, set the power and receive sensitivity to minimal distance.  The distance is easily calibrated.

The identification information is stored on the Tag. Your requirement would require only one tag for each robot and one Scanning device. As the fellow mentioned about collision avoidance, these products are based around an advanced logarithm of collision avoidance. Good luck with your project. Contact me at mflahr@identecsolutions.com
Regards:
Merv

Hi,

I am also looking for a solution on the distance measurement via the phase measurement as dirfind said. Is anybody know that if I can build a "passive reflector" that can transmit the parent signals back without processing the signals? I need conside the power consumption and side at child side.

If you use a pulsed rf device and put a triangular reflector on your target you have several ways to compute the distance.

1. Send a modulated ping

2. look for an echo

3. time gate the echo

ergo at the end you have a primitive radar.

There is a matlab/simulink model of this at

http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=2670&objectType=file

Although the numbers on this sim are way too high ie 10kw
and 3 ghz where you can casually live with 900 mhz ISM band which is cheap and slow

But if you don't want to do any design work then RFIC as suggested above may be the lowest cost.

rafe
rafeh1@yahoo.com
www.rdlabs.com

Hi
   I am planning to measure the distance of two or more objects that are moving on a field from a fixed point.  I know that RF can be used to figure out the distance of the objects but want to know if there are any company from where I can buy
any transeiver that can do the work. Also it has to be small and the problem is that I want it to measure distance upto 50m.  If anyone can help me in anyway possible I will appreciate it.


Thank You