Got Gyro?

[jdejoannis]

Originally Posted by kaszeta

We're using a gyro this year as well, it was really nice when the students rigged up a PID loop to keep the robot driving perfectly straight.

That sounds interesting. I wouldn't mind learning more about this technique.

/Jason

[kaszeta]

Originally Posted by jdejoannis

That sounds interesting. I wouldn't mind learning more about this technique.

Our robot is basically a tank drive (left side and right side wheels centered on the robot), driven by a single joystick:

1. p1_y determines the speed of the robot. Set both L and R motors to this speed.
2. p1_x adjusts the desired heading of the robot.
3. Every time through the loop, you look at your current heading, and the
desired heading, and use the error between the two in a PID control loop
to set the amount you're adding to one side and subtracting from the other
to turn the robot to correct the heading.

This allows us to correct for motor biases and different traction between left and right and drive in a straight heading.

We've also used this same approach with a omnidirectional drive system, although we've never gone to competition with it.

[Kyveck]

On 1014 we used a gyro drive very successfully last year. On the night before ship the mechanical teams accidentally estroyed the last of our encoders, so we had to scrap our encoder-PID drive. Since our drivetrain was fairly difficult to control as it was, we decided to work out something with our only remaining working sensor, the gyro.
Our algorithm was essentially like the descirption above. It took us about 45 minutes to code and tune. After it was done, our robot could easily be driven by almost anyone who wanted too. Even with the PID drive on the encoders, only a few people on the team were coordinated enough to not hit things. When the coding was done I was able to drive it on my own and chase the person who killed our last encoder around the room shooting balls at him. (That was fun.)
We got our camera working Saturday so that we can chase a green light on a cart around the school with a PI loop, and now are reimplementing the same gyro algorithm as last year to try to approach the cart without waffling side to side as much. Since the robot doesn't have encoders installed yet, the only feedback currently is from the camera.
I don't know how much experience you have programming for FRC, but if you need some help, I can talk to my programming team and get some code samples.

Kyle W
Engineering Director, Team 1014

[purduephotog]

Originally Posted by Kyveck

On 1014 we used a gyro drive very successfully last year. On the night before ship the mechanical teams accidentally estroyed the last of our encoders, so we had to scrap our encoder-PID drive. Since our drivetrain was fairly difficult to control as it was, we decided to work out something with our only remaining working sensor, the gyro.
Our algorithm was essentially like the descirption above. It took us about 45 minutes to code and tune. After it was done, our robot could easily be driven by almost anyone who wanted too. Even with the PID drive on the encoders, only a few people on the team were coordinated enough to not hit things. When the coding was done I was able to drive it on my own and chase the person who killed our last encoder around the room shooting balls at him. (That was fun.)
We got our camera working Saturday so that we can chase a green light on a cart around the school with a PI loop, and now are reimplementing the same gyro algorithm as last year to try to approach the cart without waffling side to side as much. Since the robot doesn't have encoders installed yet, the only feedback currently is from the camera.
I don't know how much experience you have programming for FRC, but if you need some help, I can talk to my programming team and get some code samples.

Kyle W
Engineering Director, Team 1014

Very interesting. Did you start with Kevin's code? We're going to be trying this this year- assuming we ever get the Gyro offering.