Chief Delphi - Sensors?!

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)

- General Forum (http://www.chiefdelphi.com/forums/forumdisplay.php?f=16)

- - Sensors?! (http://www.chiefdelphi.com/forums/showthread.php?t=153324)

Hi All,

I'm a junior from New York, team 2872 and want to know about sensors. This is my fourth year competing in FRC but we've never used and light sensors before. My teams goal this year is to be fast, small, and be gear oriented. My question is that what sensors should we use to help with aiming for the gear lift, and how to implement them. I'm our team's driver and only have experience with doing everything manually and I know that trying to finesse putting the gear on without any camera or sensor help would be a pain. I'm a noob in this area so any help is appreciated. Thanks so much and happy building!

There are vision targets on the lift peg. The specifications for the targets can be found in section 3.13 of the game manual.

Quote:

There are also vision targets on both sides of each LIFT peg. The peg targets are 2 in. (~5 cm) wide by 5
in. (~13 cm) tall rectangles located 10¾ in. (~27 cm) from the carpet and spaced 10¼ in. (~26 cm) apart
(outside dimensions).

You could also potentially use a light sensor to sense the black baseline in front of the airship.

I can't help that much, as this isn't really my thing.

But basically, there are uses for sensors.

I hope this helps at all.

You might find that using a camera and a LED light ring is beneficial in this situation.

What you can do is detect the vision targets on a peg, find the center of the target, and calculate how much you need to turn based on the image. With that information it shouldn't be too difficult to program a drivetrain mode that auto-locks onto the peg using a PID.

There is plenty of example code floating around on this subject from last year (I would take a look at the recent code postings). Also, FIRST provides a lot of example code for LabVIEW teams, so if you are one of those, you should take advantage of that.*

*That might be the same for other languages, but my team used LabVIEW, so I wouldn't know for sure.

Even if you are not up to vision (though programs like GRIP make it much more accessible) a camera feed back to the driver station would seem very valuable this year with all the stuff in the way. Start there - and work on getting vision tracking going after getting the feed established with the LED rings. (FirstChoice has them)

Regarding:
"...the vision targets on a peg... ... FIRST provides a lot of example code for LabVIEW teams"
has anyone found the Labview example code for the lift peg targets?

If you can do vision processing it's great, for others please think about effectively using pilot. How can he see spring is aligned? Will a flash light help by illuminating the spring? Can he help you align the robot? There is a green or blue tape on the carpet, that goes right to spring, but the contrast is not significant for light sensor. More over the carpet and tape color can keep changing due to dust and wear. Can pilot use this tape to help in aligning?

Sensor #1 for us is going to be the streaming camera located a few inches below peg level (so it doesn't get impaled). We will probably also have a rear streaming camera (though we only stream one at a time) for gear intakes, and a front camera to detect the vision guides for autonomous. There are also teams which use cameras to automatically find game pieces and goals. Edit: for streaming drive video, do low resolution, relatively high frame rate feed. For targeting taking place on the robot, use a bit higher resolution even if it means slower frames, and (assuming you're talking about a light ring and retroreflective tape) short exposure times. For teleop, some teams have used their drive camera, and marked the "best" location of the retroreflective tape on their laptops, either digitally through the drive station code, or by putting strips of painter's tape on the screen.

Short-range rangefinders (on a budget, I recommend infrared for ranges up to about a foot, ultrasonic at longer ranges) may be good this year, to let you know if you've got a bunch of balls between you and the wall with which you're trying to retrieve or score game pieces. Inexpensive rangefinders are available in both analog (voltage is a function of range) and digital (yes, something's there!) versions.

Encoders on wheels will improve the repeatability of autonomous driving for any drive train; many consider them essential for more advanced drive trains.

Another sensor type to consider for gear management are some IR or light-based interrupts to let us know when we've got the peg through the gear, and whether the gear is inside or clear of the robot, etc.