View Single Post
  #11   Spotlight this post!  
Unread 18-10-2011, 18:29
Dave Scheck's Avatar
Dave Scheck Dave Scheck is offline
Registered User
FRC #0111 (WildStang)
Team Role: Engineer
 
Join Date: Feb 2003
Rookie Year: 2002
Location: Arlington Heights, IL
Posts: 574
Dave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond reputeDave Scheck has a reputation beyond repute
Re: What won in 2006?

Quote:
Originally Posted by Brandon Holley View Post
I think the closest a team came to doing this was 111. Maybe some of the guys from WildStang can chime in as I am definitely just recalling this from a foggy memory.

If you think about the logistics of doing what you proposed, it ends up being quite a challenge.

-Brando
We definitely experimented with this early on. The goal wasn't to collect and shoot, but rather escape from a defender and shoot. We mounted a laser pointer to our turret and drove around while the turret tracked the goal using the camera. We found that the system wasn't fast enough to make the shots on the run at driving speed. We thought about adding in some predictive logic, but the turret wasn't fast enough. In the end without extra logic and keeping it relatively simple, it tracked great at longer distances and while sitting still. It worked pretty well when moving slowly and while getting pushed.

In auto we wouldn't shoot unless we were locked on to the light. In driver control, we had an LED stalk that indicated when the turret was locked on. This gave the driver full control.

In terms of trajectory, like many teams we had a hood that rotated to change the angle of the shot. We found that when close to the goal the angle had to be large and it decreased as we moved back. Then at some point the angle had to go up again to make the shots while the ball was coming down. We created a lookup table with the perfect angle at various distances and did a linear interpolation between the points. Using the camera, the robot calculated the distance using trig and adjusted the angle based on the lookup table. It worked pretty well for the most part.
Reply With Quote