Lasers are used alot yes, however my interest lays solely with camera work. The special obstacles I am using help the robot establish path ways. My longer-term goal is for the robot to learn its environment (using mind-maps and pattern matching), so that the obstacles can be removed.
So, the bot needs to 'see' the object from at least two locations to determine the objects position by triangulation? It's kind of like plane table surveying. Here is a link that may help
http://www.thriplowlandscape.cambscommunitygroups.org.uk/landscape%2010.htm
I ain't skeert of nuttin eh?
Yikes! What was that?