As part of our efforts to promote ROS, the Meetup group SV-ROS (Silicon Valley ROS), has been using the Neato BotVac as a learning platform (Sort of a role your own Turtlebot). I have been working with using the intensity of the returned signal from a Neato Botvac XV-11 type LDS (laser distance sensor). Recently, we have been adding intensity to our own version of the neato node to our Git Hub repository, “intro_to_ros.”
The Neato company uses the intensity to locate their charging stations. It is suspected they actually use a pattern of returned intensity data to locate specific charging stations. What I wish to do is use the intensity to 1) locate bright (highly reflective) objects. 2) actually be able to recognize 2-D markers on objects. This seems possible, and I am wondering if this hasn’t already been done.
The ultimate goal is to complete the three challenges for the Homebrew Robotics Club “Floor Bot”
1. Make a robot go from one end of a room to another and then return to its starting position.
2. Locate an object randomly placed in a room.
3. Retrieve and object randomly placed in a room and transport it to another designated location.
http://www.hbrobotics.org/robot-challanges/
I’d be happy to collaborate with someone on this. I suspect there is already a clever algorithm out there that can decide that an object is bright for its range, but I am not sure how regular light (inverse square law) compares to laser light from a reflected surface. At any rate, someone needs to figure out how laser brightness is affected with distance from retro-reflective surface.
The above chart based on only 6 points shows the intensity of the charging station is at least 2 times that of white paper.