swarm , find position of robots

Hi I'm planning for a project which involves a set of robots that can communicate wirelessly (say, using XBee or simple transceivers) .

I want one robot to find the distance(need not be accurate) between the other robot . I may use a range finder but first i must know where that particular robot is .

I don't want to use GPS modules or camera modules as they are costly or not supported aurduino boards [If there are any low cost and arduino supported above 2 modules , it would be great if you could send me a link ]

What h/w can i use ? All I need to do is identify another robot .

1 method that I thought was use relative positioning , i.e start all the robots from the same point and as the robot moves increment/decrement a (x,y) variable stored in it based on the direction it moves . However this method is very basic,inaccurate and non-scalable . i'm looking for a h/w solution to this .

thanking you .

codeRobo,

I'm not sure whether there's any off the rack solution available.
But, long long time ago (it seems long for me), TI was talking so much on the location engine on their CC2431 Zigbee chip. The software stack that time was so unstable when we were testing them back in ~2006/2007. But, I'm not sure what is the current status on the CC2431. If there's any company that develop modules (Zigbee) that is based on the CC2431, the location engine feature might be included. But, if I recall correctly, you need 2 (at least) stationery nodes (robots) in this case as reference location. If the module is available, then making an Arduino Shield should be feasible.

what accuracy do you need ? you make two contradictory statements !

I want one robot to find the distance(need not be accurate) between the other robot

AND

method that I thought was use relative positioning , i.e start all the robots from the same point and as the robot moves increment/decrement a (x,y) variable stored in it based on the direction it moves . However this method is very basic,inaccurate and non-scalable

you could use optical encoders which might not give huge errors but depends on different factors
NOTE :- accuracy and errors are relative to you so please specify
( at the begining even i didnt know this :wink: but we should know this )

sorry :slight_smile: by low accuracy i mean may be around +/- 20% error will do .

Thanks for the cc2431 location engine concept , i didn't know that , However its hard to find data about it . Any links about where to buy them and how to use them would be useful .

Where can I buy optical encoders , How do i read its data ? please send me the links as well .

Where can I buy optical encoders , How do i read its data ? please send me the links as well .

Why? Can't you use google?

by low accuracy i mean may be around +/- 20% error will do .

if so encoders will surely do the job do a google search for wheel encoders

Encoders or dead reckoning should prove accurate to 20%. The big disadvantage is that slippage won't be picked up. You might consider something like this for direction (but not distance): Pololu IR Beacon Transceiver I don't think there is a cheap solution. :-[

codeRobo,

What are you thinking of making? A friend of mine was thinking of making a simulator for bee attacks.

The cc2431 location engine concept is not that accurate when the nodes are close together. You say 20% but what in absolute distances are we talking about here.
This is something that gets asked a lot and I haven't seen a satisfactory solution yet.

let me start this clear from first....
In another thread , named , " swarm mapping "