Thoughts about future autonomous scoring...

[JimGRobot]

This thread may belong somewhere else, but I will start it here...
I sent an email to Dean Kamen, or whoever reads his mail, about this topic, but decided today that this forum would be a good place to start talking about this.

Short Version: Should we promote some way to create a separate award for autonomous scoring so we can have a serious challenge without tilting the "playing field" too much?

Long Version: Last year, robots that could score well during autonomous operation could determine the outcome of the game. The challenge was difficult, and deserved the reward of a high score, but I think this was an "unfair" advantage over the robots that could score well, just not during autonomous ops.

Contrasted with this year, the challenge was equally difficult, but hanging a Keeper had little impact on the outcome of the game. I suspect that this was by design to even out the competition a little bit, but that is only a guess.

I would like to see a way to make the autonomous ops really count for something so that more teams might give it a serious try. However, I think that it should be somewhat detatched from the scoring of the matches. One way would be to have a separate award for the high autonomous scoring robots, and not have it be part of the match scoring at all. Another would be to have one element of, or a portion of, the autonomous scoring apply to the main match.

Just something to think about during the off-season... ;-)

[Donut]

I wouldn't do an autonomous award purely based on score, as there can be some pretty cool autonomous thing done that don't score any points (for example, 190 could go hang on the bar in 2004 during autonomous, insanely difficult and cool, but technically not scoring any points).

I think it wouldn't be a bad idea, "Best Use of Autonomous Mode" or something. Some would argue that the Innovation in Control award is sort of based on autonomous usage, but I don't know how much it plays in.

[Tom Line]

They need to keep it in a state of flux. Some years will have extremely important auton, and others will have nearly pointless auton. After all, some teams have a great group of programmers, and others have a great group of machinists.

I believe this year is the year that the mechanical guys really get to shine. If you have a good mechanical group, you can have a good robot. Last year, without a top notch programmer who could do auton programming, you were not in the top tier.

[EricH]

I think this should be a poll. Personally, I'd like to see some autonomous awards. Possibly "Most Consistent" and "Coolest Auto Mode" (or something like that.) That way, the teams with good auto modes and bad driver modes get awards. (Not that there are any teams like that...I'm just coming up with an example.) Somewhere along the line, a Rookie Autonomous (for the best rookie in auto mode).

By the way, 190's auto mode in 2004 was really key to their success--50 points easily, plus their opponents couldn't get up unless they were allowed to. We did something similar, but we chose not to hang in auto mode. But yes, cool things can be done without scoring any points immediately, and maybe those should be spotlit somewhere. Hey, maybe if FIRST doesn't do auto awards, the programmers on CD could evaluate videos of cool auto modes in between debugging code and downloading it and writing some more and developing cooler auto modes...

I'm going too far here...or am I?

[Vogel648]

Unfortunately, last year in a game where programming was important we were starting fresh with brand new programmers, and this year where we can do autonomous it isn't worth that much. It can still end up being the difference this year and really gives you a morale boost if nothing else. Honestly, I think that it is difficult to make an autonomous challenge that is apropriate. I think it would be good to make it part of the regular match as it currently is, however I think that the rest of the match should be important too.
I think this year is a lot closer to balanced than last year. A bonus 5 points or something along with the autonomous ringer would have been a nice boost without overbalancing it really. The designers of the game probably thought the fact that it could not be nullified would be more important that it was.

[theycallhimtom]

My only problem with an award for autonomous mode, and some of the awards in general is that teams work specifically for the awards. From my two years I found the coolest thing about the competition is working as a team. This year we had a good programming group so our fabrication team could make some different choices. Were already thinking about what to do for next year when our fabrication will be strong and our software will be weak. Rather than the software people competing for the software award, and the builders competing for engineering awards I think that everyone working as a team to play the game is much more rewarding.

That being said an autonomous mode award would be great. Having spent many, many hours working on our autonomous mode (and that &%&#$*% camera) I know what teams put into their auto modes, and that they deserve recognition. One more award will not turn all FIRST teams into groups of award mongering students. But at some point having an award for every little part of the robot gets to be ridiculous. Certainly having more than one autonomous award would be taking it to far.

[Astronouth7303]

I think having some unofficial awards for auton would be cool, no matter how they're done. More teams need to "get it" with this whole "smart machine" thing. (There are many that are adverse to anything more to a fancy RC car, for no apparent reason. I've had to fight this all year.)

Autonomous mode is also a good chance to bluff. You load your robot on to the field with a keeper, only to find out that all you do is drive forward and put it right in front of the robot on the other side. Not necessarily a useful auton, but it has more than bluffing merit; you would be clogging the field just a little bit more.

Also, a lot of the best robots involve a lot of autonomous-like stuff in operator mode. Instead of hooking joysticks to motors, you take joysticks as input (more like suggestions) and drive control algorithms based on it. Meaning: Use same/similar code for autonomous and drivers. I know one team won an award at the Buckeye regional back in 2004 for there gyro-straight user mode. (Meaning, a gyro kept them driving straight, even in user mode.)

The easiest way to write autonomous IMHO is to do the hardware-interfacing legwork in advance and write a particular autonomous using that code.

[bear24rw]

Alot of our teams robot functions were autonomous.. the whole arm was pretty much autonomous, press a button for which height and it rasied the tower, extended the arm and set the wrist.. it was all handled by 2 functions which we used in both user and auto mode. We also had dynamic braking where if you held the four buttons on the drive joysticks the RC would put the brake on the speed controller and it would try to keep its place with the encoders on the wheels, so that was autonomous, again we used the same function for automode.

I agree with others, autonomous mode is something that seems attempted by few. This years was pretty hard, and I understand that if you were a ramp only there wasnt really alot you could do, but I would like to see more teams focus on it and trying to make it a priority for their team in the future.

[ericand]

Originally Posted by bear24rw

Alot of our teams robot functions were autonomous.. the whole arm was pretty much autonomous, press a button for which height and it rasied the tower, extended the arm and set the wrist.. it was all handled by 2 functions which we used in both user and auto mode. We also had dynamic braking where if you held the four buttons on the drive joysticks the RC would put the brake on the speed controller and it would try to keep its place with the encoders on the wheels, so that was autonomous, again we used the same function for automode.

I agree with others, autonomous mode is something that seems attempted by few. This years was pretty hard, and I understand that if you were a ramp only there wasnt really alot you could do, but I would like to see more teams focus on it and trying to make it a priority for their team in the future.

The whole concept of autonomous vs automatic functions was hard for the team in general to understand. We used a bunch of automatic function (to raise the gripper to a hight, to score when we were in position), but we could do more. The problem is often one of terminology and understanding.

Once the team understood how automatic functions (during tele-operated mode) could speed up our average time to score, we got a lot more support. Before the concept was fully explained, we got some resistance from people who equated working on automatic functions with autonomous mode and who wanted us to be focusing first on driving the robot.

I think a good set of automatic functions make control during tele-operated mode much easier. They also provide well tested building blocks for autonomous mode functions.

[Brandon Holley]

One reason autonomous isn't as big this year is because less people are going for it.

If no one is scoring keepers, than its not really a big deal if your team doesn't score keepers...but if say 2 out of 3 teams are scoring keepers, then your team might start reconsidering scoring them, EVEN IF they don't mean that much to your overall score....

Just something to think about...

[ldeffenb]

Originally Posted by Brandon Holley

...EVEN IF they don't mean that much to your overall score....

I beg to differ that autonomous doesn't score much. Rather than considering it to be "just 2 points", consider instead that it "DOUBLEs what I score adjacent to it"!

Lynn (D) - Team Voltage 386 - Winner of 2 Awards for 14 Autonomous Modes

Orlando Regional Xerox Creativity Award
Palmetto Regional Innovation In Controls Award

[efoote868]

Personally, I would like to see a longer, more meaningful autonomous period, perhaps 20-30 seconds, along with a better goal. This year, we programmed an autonomous, but once our robot was in perfect condition, the team didn't want to see us breaking the robot (with several matches, other robots got entangled with the rack, and spent the rest of the match untangling). Hopefully I'll be able to convince them to use it at IRI...

This year, the autonomous (IMHO) looked lucrative, if you got it right, you could place a spoiler on your opponent's side, making it nearly impossible for them to get easy rows. Its just our programmers never got the robot until near the end of week 4, and even then we had to wire, re-wire, and re-re-wire (ending in a nightmarish tangled mess).

Although we didn't get to use autonomous, we did however make the robot much more like a robot... whenever our antlers were empty, it would turn on its kicker wheel, and be ready to pick up tubes... when it got a tube, it automatically went to the previous height specified... when it got near the rack, it popped the tube onto it. Next year we'll do that... and hopefully for years to come we'll do that.

[gblake]

Originally Posted by efoote868

Personally, I would like to see a longer, more meaningful autonomous period, perhaps 20-30 seconds, along with a better goal.

Agreed.

One thought that I intend to start working into Potomac Vex League scrimmages is giving robots an uncontested field for X seconds and seeing who can score the most points in that time.

Insert the word autonomously into the paragraph above, squeeze the action into the breaks between regular matches or do it during an evening (sort of like the home run derby that occurs outside of baseball's All-Star game) (or install a second (half of a?) field at each regional (Yikes!)), and you have decent reason to give out an award called "Autonomous High Score".

Blake