Quote:
Originally Posted by Stud Man Dan
Lets make Hybrid exciting, teams flying around the field scoring dozens of points narrowly avoiding disasterous crashes.
...
Take their challenge, a negative attitude towards it won't get anyone anywhere, but a positive outlook on hybrid may produce some spectacular performances, and I applaud the teams who try.
|
Furthering one quadrant for the short distance movers furthers the long distance movers by at least one, perhaps two. The majority of the time against strong hybrid players, net gain for short movers is, at best, zero.
Positive, negative, or neutral.....there are going to be a lot of teams who will end their autons, voluntarily or otherwise, one quadrant further down the pike than they had ended them previously. Except now, instead of meeting these obstacles at a relatively slow speed as they made the lane change, oncoming hybrid modes of those successful enough to program multi line crossings at rapid speeds (we will be one such team, although we only program 60% of full speed currently) will be contacting many of these obstacles after having half a field of acceleration time. You cannot possibly guarantee that 100% of hybrid routines will "narrowly avoid disasterous crashes". Bad luck will dictate that some will not be avoided. I'd have preferred retaining the low-speed collisions of Week 1 but now we must all be concerned with this new risk. So again, I caution, be careful. Dare I even suggest, slow down.
Driving (relatively) straight forward and stopping is basic enough to implement for teams who have never done it before, but there are still roadblocks to meeting this goal. I can realistically see this as a solution for most teams who do not currently have a hybrid mode. I've got code handy to help them with the task. But there are still roadblocks to this effort that must be overcome at events. I attempted to help one rookie team out at Midwest, but their programming mentor was absent at the time, and I wasn't going to commandeer their laptop and force feed it to them. I left the code in a text file with my phone number and encouraged them to call me later on when they had time to receive further assistance. Heck, some rookies don't even have access to a laptop. I don't feel putting the code in there is worth the time and effort if I can't realistically explain it to the kids and get them to understand it. In the short amount of time we have at an event, this is a true roadblock.
Driving forward, turning a timed 90 degrees left, and changing lanes is much more of a challenge than even driving straight. A lot of repetitive trial and error is involved when implementing timed turns and moves of varying distance. Many teams are more concerned with passing inspection and being able to drive at all during practice matches than they are getting a 2-3 line autonomous to run. At some venues, there is no room for a team to tether up somewhere and verify how far their timer settings make them go and turn. So what happens when those teams program too long of a turn, point their machines CLOCKWISE down the field, and let er rip? What happens when delays are not sufficiently programmed between alliance partners, and collisions between teammates result in robots driving the wrong way, increasing their risk of damage? Are we all comfortable with such risks? That is a true risk of any hasty hybrid programming push.
One thing that aided us in autonomous tuning was videotaping our practice matches. Based on the replay timings we viewed in our pit, we were able to extrapolate the correct timed values we needed for fairly consistent turns, and a 2-3 line auton resulted. I can offer such a service to anyone who comes asking for such assistance in Pittsburgh and Cleveland, but I will not look to shove hybrid code down the throats of those who either do not want it or are not ready for it.
The build season and offseason are the TRUE times at which hybrid programming education should be promoted. With access to teleconferencing and remote desktop software, there are ample opportunities for veteran teams to reach out to long distance teams who wish to receive autonomous assistance, and there is enough TIME available to all to ensure that the resulting code is functional and safe. Letting an inexperienced team unleash an untested hybrid mode on the field at a qualifying match is a haphazard and careless exercise, in my opinion.
Be. Careful. Let's all merge these idealistic wishes with the true realities of the event venues, then proceed accordingly.