Difference between revisions of "Robot Sensor Horn for Sharp IR Sensors (Q43410)"

I designed this sensor horn in order to implement a very simple mapping algorithm on a mobile robot. This allows you to see all 360 with two sensors and a standard micro servo(I got mine in a ten pack off ebay for not much at all) in one 180 degree rotation. I had this made out of wood originally. It wasn't very good. Now that I have a ToM, I'm excited to improve upon my design. I'll mark this complete when I get it to print without having to use my dremel drill press. I also need to replace the belt on my ABP, I rigged it with a layer of kapton tape for this print, but it still ended up a little warped and weird. Then there's a mountain of things I could do to tune my printer as well. :D As for the algorithm, I got around not having an encoder by timing the amount it took for the horn to make one full revolution*. (the el cheapo servos only move at one speed) I take the total time it rotates and divide that by the amount of samples it took during that time for one sensor. I then multiply that number by 180 degrees to get the increment of the theta component for each sample. The end result is a set of polar coordinates for 360degrees around the robot. You could also just take a measurement every (rotation time)/180 for each sensor, which gives you a smaller array, but it's a bit easier to work with. I didn't really see too much difference. *the robot still moves the sensor past this time to make sure it reaches the maximum of its travel. The robot just waited a set delay when moving one direction before taking measurements (since only one side traveled too far on the little servos I have). Its a bit inaccurate, but it worked well enough to hug walls, avoid chair legs, and go through doors.