View Full Version : Future Autonomous Mode
05-01-2011, 05:46 PM
Autonomous Mode has been pretty standard in the past. At the beginning of each game robot's would start in known positions and would perform a task in a given time frame. What if autonomous became more random in future games? Maybe autonomous mode could occur in the middle of each game (teleop, autonomous, then teleop again) so that a starting location wouldn't be guaranteed each time. Or maybe autonomous would occur at a random time within the game. This would promote more dynamic autonomous modes instead of dead reckoning. What do you think? What would you like to see out of autonomous mode in the future?
05-01-2011, 09:09 PM
Your post has started me thinking (congrats on a successful post :) )
I think it would interesting to have an "autonomous zone" on the field, perhaps surrounding the place where the robots have to go to score or to acquire game pieces, wherein the robots have to operate fully autonomously. There's always the concern raised of making the game accessible to rookie teams/teams with limited resources, so perhaps there would be lower-point scoring opportunities elsewhere on the field, but in order to get the "big points," autonomous functionality would be required.
I think it will be better to have the 30sec end game as the auto. mode. That will make things more interesting
05-02-2011, 06:39 AM
See i've always been a proponent of the mid game autonomous. Personally how I'd like to see it handled is as not a forced autonomous but an optional autonomous. In other words there would most likely be no times where most of the robots are just sitting or slamming into walls as we experience in the first 15 seconds now. Rather by creating a "dead zone" with a task that can only be completed autonomously at any point in the match, you create stronger looking robots. One idea of this would be making the lanes from this year the dead zones, so once you were in them your robot could drive back by itself to get a 3x tube.
With cameras on many robots now my original idea from years ago probably wont work. My idea was to have goals facing the opposite direction of the drivers, so they couldn't see it. and make those goals so they could only be scored on autonomously. However now with cameras, you could still do it teleop.
05-03-2011, 11:59 AM
I've said this many times before, and I'll say it again that there should be an autonomous period with 3 possible objectives.
Objective 1 - 5 points
Something easy something a rookie team could do if they worked at it.
Objective 2 - 10 points
Something moderately challenging that an experienced team could do with some effort.
Objective 3 - 15 points
Something very challenging that even a veteran team would have to reach to achieve. It can't be scripted, or played back, it would truly be a dynamic ever changing every round challenge.
Why you might ask?
The easy challenge can allow the rookies to "feel good" and inspire them to reach higher.
The moderate challenge that most teams will attempt but give them the drive to reach higher.
The extreme challenge that very few will reach, but this is to "show off" if you will, their excellent skills.
05-03-2011, 12:12 PM
Logomotion struck a good balance between difficulty and reward as far as autonomous mode goes. It was fairly easy to get something that would score a tube most of the time (manipulator and lift design permitting), and FIRST provided an out of the box solution for line following. But in order to be flexible with where you start, get the reliability up to 100%, and do things like two tubes, teams had plenty of room (and incentive) to go above and beyond.
My only complaint is that with static fields, game pieces that the team is allowed to position themselves, and no contact during autonomous mode, many teams (my own included) have basically been able to do nothing but dead reckoning with a gyro and encoders and succeed the past two seasons. FIRST keeps giving us sensors in the kit, but not a lot of incentive to use them. There has got to be a balance between the ease of dead reckoning and the risk/reward of using vision or other sensors that is better than what we had in 2010 and 2011. I think 2006 was the last time they got it right.
(Yes, I know teams X and Y used cameras and line followers with great success, but the most impressive auto modes of the past several years - the 2008 1+ lappers, the 2010 bump crossers, the 2011 double tubers - have largely used encoders and gyros).
vBulletin® v3.6.4, Copyright ©2000-2014, Jelsoft Enterprises Ltd.