Kansas City Regional

[DustinB_3]

I was wondering how many teams that are going to the Kansas City regional got the camera to work and can use it to score a keeper in autonomous?

[railerobotics]

Calling all Kansas City Teams!!!

www.team935.info

[Birdman2388]

Northmen Robotics at Oak Park will be there. Because of the bad weather this year, we were not able to get the camera working. We plan on using most of our pit time to get the camera working.

[Martino]

I will be there with team 1825, We have the camera on our robot and we will use it in autonomous to score a keeper. You can see our robot here.

[WesleyC]

I'm the programmer for 1825--if you have questions on getting the CMU cam working, I might be able to help.

The weather around here has been truly horrible! Especially for us, as a homeschool team--we are forced to drive from all over the place, some members traveling as much as 45 minutes one-way to get to a meeting.

However, yes, we did get our camera working. The lens is off-center, the default pan/tilt values were useless, and some of the numbers needed tweaked--but after a lot of effort we did eventually get it working quite nicely. However, our testing platform didn't allow for us to test very thoroughly with multiple lights being in view--I have a tweak in mind for when we get there, but I can't do too much with it until we actually get to a real arena.

[DustinB_3]

I have actually gotten the camera to track just fine. I am just having problems using our gyro to turn to the angle of the light. The angle is so minute that the robot either doesn't move or it undershoots. If you have any advice please do tell.

Dustin

[WesleyC]

The key here is to use a PD/PID loop in combination with the PAN_SERVO value provided by the camera's tracking code. Set the gain large enough, and even small differences in the camera's left/right position will cause your robot to move.

[DustinB_3]

I have set the gain to 500 and the robot sits there and you can here the motors trying to move the robot. So I set the I higher but then it overshoots. Since it's in a case statement I am thinking I need more or less of a range for it to go to the next case. I have been working with it for 3 days straight but it still wont go to the correct angle.

[WesleyC]

Why use a case statement in a PD loop?

The gain I used on our robot is 5 or 6... however, the output of the PD loop is then sent to our drive code, which incorporates an exponential drive algorithm.

The best thing to do is to use a printf statement right before the drive motors (or, even better, AFTER) are set--this will enable you to read EXACTLY what the drive motors are getting.

Since this may not be possible with the robots having been already shipped, I'll give you a quick fix:

By now, since you've probably figured out a minimum value that DOES make the motors move the robot, set a MINIMUM speed limit. Something like...

//150 and 105 are just arbitrary numbers; replace them with the minimum speeds that turns your robot

//copy this code for each motor
if (motor_speed > 129 && motor_speed < 150)
{
motor_speed = 150;
}
else if (motor_speed < 126 && motor_speed > 105)
{
motor_speed = 105;
}

Do that for both motors, and it should limit their minimum speed. If they're getting a value a little off center it'll increase their speed to the value you define as a minimum. That stops them from "humming" and actually makes the robot move, even at very low speeds as defined by the PD control.

[DustinB_3]

I'll try that. Good luck it Kansas City.

[WesleyC]

Hope that works for you!

Good luck, and we'll see you in March.

[kawelch]

One week to Kansas City's Inaugural Regional. Good luck to all the teams attending. Let's make it a GREAT one.

[Barry Bonzack]

Looking for a webcast so the community of team 2165 can watch the team during their first regional. Does anyone know of the URL where I can find this?

[DustinB_3]

The URL is:

http://www.more.net/services/videost...trobotics.html

[Sugar_Bunny]

How are things shaping up?