View Single Post
  #11   Spotlight this post!  
Unread 18-06-2011, 04:47
Tristan Lall's Avatar
Tristan Lall Tristan Lall is offline
Registered User
FRC #0188 (Woburn Robotics)
 
Join Date: Aug 2001
Rookie Year: 1999
Location: Toronto, ON
Posts: 2,484
Tristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond repute
Re: Driver Selection: A Discussion

In theory the ideal driver doesn't even need a coach. They're so masterful that they have the strategy covered as well. Good luck finding or creating that driver.

I'm not sure if I was trying to pull that stunt when I was driving a robot, years ago. We had a combination of a first-year university student mentor as coach, and me as driver—and although we got along excellently, I'm really not sure how accurately I reacted to his coaching suggestions. (Then again, maybe he was so good at planting suggestions that he didn't need to explain anything. I really wasn't consciously paying attention to him...which is something I wouldn't advise under ordinary circumstances.) Most of the time it worked, but there were a couple of tactically-excellent, but strategically-unfortunate decisions that would have benefitted from a more refined coach-driver relationship (like pushing an opponent's robot 50 feet across the field, including up and over a big ramp, kicking and screaming...and getting pushed most of the way back when once our battery voltage dropped low enough to reboot the robot controller ).

An interesting side effect of that particular robot was that the function operator (i.e. driver #2) had very little to do. From a controllability point of view, it probably would have been better just to give full control to one driver. But on the team at the time, being a driver was also a prestige position, and that fact (combined with some implementation details of the controls) made it impractical to design the interface around a single driver and leave the other driver with literally nothing to drive.

Desirable though it might be to share the prestige and responsibility among two people, I think a team needs to be ready to hand off complete robot control to a single person, so that there's one less brain in the loop trying to make the robot do stuff, and one more set of eyes watching the field and manipulating game pieces (when possible).

And if the drivers can't calculate strategy on the fly as well as a dedicated coach (and especially given the calibre of coaching available to some teams, the drivers very often can't), then the coach needs to be assertive enough to be noticed, competent enough to understand each game on the fly, and consistent enough to avoid ambiguity and communicate nuances seamlessly. Practice makes this work. (In my days as a student, we never had enough time to practice with a working robot.)

Similarly, making the coach a prestige position doesn't really work. I tried coaching once; I wasn't great at it. Firstly, I was coaching my brother as driver—so what were the odds he'd listen to me unquestioningly? Secondly, I'd been primarily involved with the technical development of the robot—so I was always on call to help fix robot issues. We didn't really have a dedicated strategy team at the time, so I didn't have the benefit of other peoples' research into what the opponents were doing, or even the background knowledge to choose what the strategy should have been for the next match.

The coach and the drivers need to be on top of the strategy from the start. That's what they'll be applying throughout the match, so it makes sense for them to dedicate their energies at events to that task alone. Don't let the pit crew drive. You need them to make sure the robot is functional at all times. Despite that rigid separation of roles, there needs to be some cross-training of skills. The pit crew needs to understand strategy to prioritize repairs—what can we do without next match? The drivers and coach need to understand robot operation at a technical level, so that they can perceive malfunctions, and don't try to make the robot do unrealistic things.

Rewarding excellence is important, and since your most skilled students probably deserve recognition for that attribute, you'd make them very happy by letting them take the prestigious role of driver. So in a roundabout way, driving can be a prestige position without harming performance. But that only works if the all-around skill is there—mastery of strategy by early in build season, mastery of driving toward the end of the build season, and competence in terms of understanding robot performance and failure modes.
Reply With Quote