Has anyone else had trouble staying within the 10 second limit in autonomous?

Has anyone else found that 10 seconds is not enough time to move the robot to the light? I am using the camera to move the robot to the light. The fastest time we have had to get to the right spot an start shooting balls is 9 seconds. Obviously, 1 second is not enoght time to shoot balls, and also, most of the time it takes at least 15 seconds to get to the right spot. I decided that rather than going to the same spot every time, to have a very big dead-spot for tilt and then change the cannon speed(our cannon is fixed) based on exactly where we are. I have messed with the GAIN and even made my own system for it, and I have concluded that there is really no way to get to any spot in under 5 seconds. Sometimes it may take 5 seconds just to lock on to the light. If the GAIN is too high, it will keep over-shooting, and if it is too low, it is too slow. The perfect number is still to slow. Has anyone been more successful than me? Do you have any ideas for me? Is it possible that blob size and those other values would help?
-Dave
-Longwood team 564
[email protected]

Right now, I’d say 5 seconds to reliably lock onto the light. But you can be searching while you are getting in position. It is possible to fire 10 balls in 5 seconds. Probably even in 3 seconds. But I agree, I’d have rather seen the traditional 15 second autonomous.

Ok, good, I didn’t realize that you could be locked on before autonomous starts, that cuts off a lot of time. But still, you will be facing straight, you may lock onto the wrong target, and even if you do happen to lock onto the right target, you have to make a turn twords the right target, and during that turn, it make loose the target. Thank you for the amazingly fast response at 3:30 in the morning. Clearly we have no lives, lol.

I don’t think 10 seconds is enough to do alot, but for myself, I’m hoping to at least move the robot to firing position.

I still haven’t been able to mess with auto mode but is it really that hard to get everything in 10 secs?
Also is there any other purpose for the camera outside of auto mode, I was thinking if I could get a led rigged up, or maybe one of those pwm01/pwm02 leds on that one…thingy…I don’t remember what its called, the one that you plug in the controllers to. But I was thinking to hook up a led so that when it was lined up to the goal it could light up so we know when to shoot.
Also I’m dead reckoning, I have a feeling I will want to cause pain to something/someone afterwords… :slight_smile:
I should try a pot, but I just don’t think we have enoguh time left me me to learn how to.

Does anyone actually plan to shoot in autonomous?

why do you think you have to be facing straight ahead when auton mode starts?

in the past you were able to point your robot any direction you wanted as long as it was within the defined starting area of the field.

Also, what if your team ends up going on defense? Keep that in mind, you might not always be able to shoot right away after the autonomous round

I think many teams plan on shooting in autonomus mode as the base to their entire stratagy. And I think they will do it, too.

But I think only one robot will be able to unload 10 balls into the center goal per match… better have a few other auto modes ready to go!!!

uh… I gotta go talk to our programmers…

I had no idea! Thanks for the correction. That applies to this year too? That would be awesome! I have to check the dimensions of our robot to see if we could do that. Also, don’t you think that the robot next to you might get in the line of sight? If they don’t move in autonomous, this could be a fatal problem for autonomous.

It could be done something like this


#define BEGIN_DISABLE     0

extern char robot_state=0;

Process_Data_From_Master_uP{
	switch(robot_state)
	{
		case: BEGIN_DISABLE
		{
			Tracking_Stuff();
                        break;
		}
	
		
		default: break;
	}

	Other Stuff...
	
	Default Routine;
}

(I put the switch statement in just in case you wanted other robot ‘modes’)


User_Autonmous_Mode
{
	robot_state=1;
	
	Autonomous Stuff;
}

Has anyone actually tried tracking while disabled?

Kevin’s code uses the PWM outputs which output neutral while disabled.

Can Kevin’s code command servo values through the camera? That is, can you plug the servos into the camera itself, and still control where they point?

Not easily.

So far as I can tell, once it sets up the camera, it basically just “listens” to the T packets coming back. There doesn’t seem to be an existing state machine to patch into for sending SV commands when desired. Changing the code to use the camera for servo control means adding a new piece of processing rather than adapting an existing one.

So I assume that means that by default the camera will not start searching until autonomous starts. With this code, will it remain locked on when autonomous starts, or will it just restart the search?

I am also thinking of using the camera or dead reckoning to get the corner goals in autonomous but I think that would take even longer! What are some of the other ideas for autonomous out there?

To help with the whole searching problem, we’re going to try to initialize the camera so it will automatically turn to face about where the target is, then search from there. That saves a second or two of sweeping around looking for the light.

The ability to start facing any direction is what really helps this year. Since we shoot out of the side of our robot, we are going to start with our side facing the front of the field, and simply drive backwards (which will be left because of the robot orientation) a bit while everything locks onto target and the wheel gets up to speed, then fire as fast as possible.

We’ll see how well auto firing works. I’d be worried about defensive bots purposely ramming you, except that you must specify which zone to start in BEFORE the match, which means they can’t move their robot to try to hit yours. They can use a switch to pick a different mode and turn to come after you though.

And a last note, why do you need to move to shoot? The center of the three starting positions are 27, 30, and 34 feet from the goal.

plan yes, but we will see how that works out

I feel that that is too far away. It is possible that we will shoot from the starting position, but we haven’t tested our new robot yet (I am working and testing on last year’s robot, which has the same chassis design) so I am not sure how good our cannon will be. It’s a good idea, but I think a: it may be too far; and b: the robot next to you might get in the way. I will definitely consider that plan.

-Thank you all for the suggestions! Keep them coming.
-David Mazza
-Longwood Team 564 programmer

I would be interested to see teams that have had a chance to test this, but it seems to me that it would be really difficult to make it from anything but the front starting position at 27 feet from the goal. If you go with the 12m/s in FIRST’s rule, and their discussion on being able to shoot 32 feet at the fastest speed, you really shouldnt be able to hit the goal at 30 feet away. But again, we keep getting conflicting answers with our testing and physics, so I would reserve judgement until I see a test that verifies a ball shooting at 12m/s makes it a certain distance.