Back to on topic:
One thing to consider is that it generally takes a lot of practice or talent to be good at controlling a robot. Think about it from a video game perspective:
Video game skill tends to follow an “S” curve relative to time put in.
Generally the first few times you try something, you are trying to understand the controls, what does what, and how the system will react to inputs. That is one of the reasons some games have skill training sessions at the begininning of them typically 10-15 minutes to complete. The games tend to ramp up slowly. Then between 1-10 hrs, there is a rapid growth of scales until the user becomes quite proficient. After 10+ hours, you tend to level off, but many will still continue to improve. The same is true of other games and other skills. Mastery level though of a skill tends to be on the order of 1,000-10,000 hours of dedicated practice.
I gave up playing first person shooter games with most of my friends as they were in the 100-1000 hour camp when I was in teh 1-10 minute camp.
The bad news is, you will not likely get there in FIRST Robotics. Even the Michigan teams that do 2 districts, the MSC, and the Championship will at most see 44 qualifying matches and 42 elim matches (this would be 3 matching every round all the way through einstein finals). This is 86 matches or about 172 minutes of controlling your robot. Thus a little under 3 total hours of tele-op mode.
Because of this keeping a consistent driver, operator, human player is often key to having a strong competitive team.
Off-loading the required skills the driver/operator need is also very important.**
Whatever practice you can give your drive team is essential. If you do not have time or resources for a second bot, “finish” as early as possible. 2 hours of *** dedicated practice ***the day before ship day is more driving than many teams will see all season.
*A not so quick word on dedicated practice. Dedicated practice is not messing around with the robot for a couple of hours. We have only had a practice bot a couple of times, but they were wonderful things to have. Messing around is fine for the driver for 1 -2 batteries as they are trying to figure out the controls. After that, you should change over to drills. Practice key strategies and techniques.
In 2007, our drive team practiced a hang 9 in 2 drill. The team would hang 9 tubes on 3 rows of 3 in 2 minutes. Initially it took much longer that the 2 minutes (they could hang 4-5 in 2 minutes initially), but eventually they got to where they could hang all 9 tubes in 2 minutes. During the matches, they typically would only hang between 4-6 because of defense, and deploying ramps, but they had the ability to hang 9 and their timing was really commendable.
In 2010, we had a practice bot as well and we spent a Saturday cleaning the home zone. The assumptions was that there were 6 balls in the home zone after autonomous, and our bot was in the mid zone. They had to cross the bump and clean up all 6 balls plus hang in 2 minutes. The first couple of times, it took over 3 minutes, but eventually they got it down to 1:20 that morning. Again with defense, the results varried, but the practice clearly improved the level of performance that the drive team had.
In 2009, we kept very accurate data on human players and could tell when teams switched human players. This became a major factor in their ranking on our pick-list.
The dedicated practice principles are for any skilled base initiatives: sports, making things, brainstorming ideas, singing, writing, or racing a car.
**Some more not so quick words on the secondary topic. Jim Z. taught me the first year I helped there team that due to the lack of practice, off-loading skills & talent based needs from the drive-team is imperitive. This would include using a gyro to help the driver drive straight at high speeds. Automatic shifting to off-load needs of the driver. Using “State” control for complicated things like multi-jointed arms that have to score at different levels. Target tracking to assist the operator scoring balls. All of this is fine and dandy, but is only possible with a good mechanical foundation.
As Jim says, great controls can make a good mechanical bot great, but a bad mechanical bot will be bad no matter how good the software is.
Last year, in auto-mode, you had the opportunity to score at least 3 balls in autonomous. The national average per team score per match was 1 pt. 1 pt!!! You could be 3x better than the national average if your team just consistently scored their 3 in auto-mode! If you were a little more daring, there was time and opportunity to score as many as 5-balls in automode. If you could score all 5 of those, you would have been one of the top 25 or so in the nation within the first 15 seconds. Think about that!
I bring this up because I know that many good teams tried to do this. I know we did. A few teams were even successful at kicking 5 into the home zone a few times. For the most part though, this was extremely limited due to 2 major factors. 1. repeatability (field, mechanical design, ball inflation, ball placement, robot starting position…) 2. Partners (not all partners will let you kick “their” 2 balls. Some won’t even get out of the way of your 3!).
), Can you really trust your TEAM’S robot’s sensors and their accuracy? What if changes are made to the robot’s physical model?