[Dprglist] LF Challenge questions

John Swindle swindle at compuserve.com
Sat Jan 13 19:55:46 PST 2018


Doug et. al,
 
A few LF Challenge questions:
 
 
1) Undershoot recovery? I understand that an undershoot is an immediate disqualification. That's a good, strong rule. Keeps one from just driving from Start to Finish in a circle. And it's more than that. But...
 
With an overshoot, the robot is allowed to re-acquire the line and try the maneuver again. If the robot undershoots, is there any option for recovery, thus avoiding the instant disqualification?
 
 
2) How close to the line? I know there are judges, and their decisions are what matter, but can we say up front how closely the robot has to follow the line? For example (and don't get too stuck on this example), if the robot passes an S-gap dash, treating it as noise or as a regular gap (thus undershooting, an immediate disqual) and then turns and re-acquires that dash in the S-gap, what should the judges conclude? In other words, if the robot jerkily marches across the dashes, missing and re-acquiring each dash, instead of smoothly connecting the dots, does the behavior match what is desired?
 
 
3) Are the lines on top of the stains? I discussed this with Ron and Will today and they are pretty sure the lines are on top of the stains. I've been assuming the worse case that the stains are opaque and obliterate the lines. Having the lines on top of the stains is easier. Can I assume the lines are on top, which means the line colors are not affected by the stains?
 
 
4) Overshoot recovery? I assume that if the robot overshoots and eventually gets back to the position and orientation where it overshot, then the judges would credit it with recovering from the overshoot and would allow it to continue (time permitting). Just clarifying this.
 
 
Interesting stuff. Good to see you all.
 
John Swindle
 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.dprg.org/pipermail/dprglist-dprg.org/attachments/20180113/6d5fdfdb/attachment.html>


More information about the DPRGlist mailing list