So my team’s looking for sensors to differentiate between the polycarb defense and the green carpet for use in autonomous. We’re already in possession of some DM7-0N-1A photoelectric sensors. How well would these work for differentiating between the defenses and the carpet?
Another idea we had would be using some type of RGB sensor to pick up the green carpet only. Does anyone have a link to some kind of sensor that would work for this?
We’d likely be mounting the sensors a bit up into the robot so they can’t get knocked around - this means that they’d have to have a decent range to always be picking up what’s below them.
Is your team planning on using some type of sensor to differentiate between the defenses and the carpet? If so, please share
I haven’t chatted with them about it lately, but our programmers were planning to try the accelerometer. When the robot’s level again (plus another few inches counted by encoders), we’re off the defense. We’ll obviously have to have some sort of FSM system to evaluate yes, we’re going up, now we’re going over, now we’re going down, and now we’re across.
The visual rangefinder to the goals (camera + target processing) can also help determine if you’re still in the outer works or not.
Your best bet might be an accelerometer. Use it to sense the direction of gravity - if your on the defense, you’ll likely be tilted in one direction or another (except for that split second your at the top of the defense driving over it).
A photoelectric sensor might work, too - shine a light down, and measure how much is returned. You might find there’s a big enough different between carpet and the defense, but that would definitely require some on-field calibration.
The field line sensors (not sure of the actual name) from the 2011 kop will be able to differentiate between the defense and the carpet. We have used then in previous years and they are quite useful (except for when the GDC puts black tape on grey carpet) Based off your team number, you probably still have them in some dusty corner. If not, we’d be happy to lend you one or two.
We have had success so far being able to differentiate between the green gaffers tape and the green carpet using it, so it should also be able to tell whether or not your are on a defense as well.
Yep that’s them. Last year we tried using two of them on the sides of our chassis to automatically line up on the scoring platform and drop a stack. It didn’t work out for other unrelated reasons.
We have decided not to use them, as they need to be close to the ground to work well, and we are using pneumatic wheels.
What kind of drivetrain are you using?
How far off the ground are you having yours?
We haven’t tried it on our actual bot yet, but testing with just the sensor in different lighting conditions showed it was reliable to ~2.5" from the carpet/tape, which is around the height I expect to have ours mounted. This was with only rough tuning of the adjustment dial on the top, so it can certainly use more experimentation.
We are also using pneumatic wheels, and after measuring it looks like we could get them as close as 1.5 inches, but more likely if we wanted to keep the sensor protected it would have to be higher than that.
With some additional adjustments I was able to get the detection working significantly further away, though I still have not done it on the actual robot. (Hopefully that will happen in the next week.)
I tried doing this in both dark and bright environments with the big ring lights, but I did not observe any difference between the two with or without the light.
