Quote:
Originally Posted by XXShadowXX
I was hoping that we would mount the camera far enough back, so that it would cover the majority of the field we could fire in. Position would be detected by change in said objects size. Movement would be calculated after 2 frames, and acceleration in 3 frames. Naturally the later info for acceleration would allow us to figure where the trailer would be in say a couple tenths of a second that it would take the ball to get there, and for a second ball to get there so we could lead enough for say 7 or so balls to reach the trailer. Oh lots of trig to do this. The cRIO can do floating point right?
|
Due to the trajectory of the balls, I would estimate half a second to reach a trailer. Unfortunately, this allegedly 30 fps camera seems to have been reduced to 10 fps (frames per second), and with horrible blurring.
Labview has a standard of 64-bit floating point integers (dbl), but there is an option of 32-bit floating point integers (sgl).
I don't believe it is legal to attach anything to the trailer or trailer hitch, as this would likely be outside the allowable limits for the robot. Assuming the trailer hitch tongue is steel, you could possibly use a very strong neodymium magnet on a spring-loaded potentiometer to detect which side it is on. More likely, though, would be looking down the tongue and dealing with reflectivity. I'm rather surprised FIRST didn't allow a reliable way for us to detect this.