For those teams who have already done both, which has been faster? The autonomous movement along the three paths, or the drivers movement along the same paths?
I’m curious how much of a difference there has been between the two challenges on the same routes. As well, the differences for specific routes, if there are any.
Which has a faster time overall?
- Both are approximately similar
- Haven’t preformed both
Haven’t done ours yet, but my hunch is that autonav will be faster (though it depends heavily on the teams autonomous capabilities and their driver skill). For teams with solid autonomous, they can do essentially as fast as they want without a major reduction in speed, as well as not having to readjust from minor mistakes. Autonomous routines are GREAT at doing pre-set movements really consistently, while drivers are better at strategy and adjusting on the fly. However, with enough practice my bet is a good driver get pretty darn close.
In general our programmers are beating our divers, but just barely.
Currently our AutoNav is way faster. Especially on Bounce.
We haven’t really tried driving the paths manually yet, (We should probably get on that) but I have a hard time envisioning our drivers going faster than the autonav. Just like there weren’t very many (any?) manually driven 2 hatch sandstorms in 2019.
There were a ton of human driven 2 hatch sandstorms in 2019, the biggest two examples I can think of are 3538 and 1619 (3538 didn’t even have vision tracking!).
EDIT: and as I have been reminded, the OP god Karan on Shaker
Didn’t 973 run the whole season without an auto as well?
Afaik 3707 and 20 did as well. Anyway, yeah, there were a ton. Just required a bit of practice. Maybe nearly as many as autonomous 2 hatch sandstorms lol.
Currently we have got in a decent amount of time for both hyperdrive and autonav. the driver(me) has been driving the three paths around 2-3 seconds faster than the autonomous. We hope to speed up the autonav with more tuning but we dont forsee the software ever being faster than the driver.
I think autonav was faster on the one path we had programmed (barrel run), but then we took our ‘hanging’ mech off the robot, and had to make some adjustments to both auto and driving. Drivers were slower today on all paths, but auto was readjusted and almost back up to speed. My prediction is that when we’re done, AutoNav will be faster on all paths. But we do have a couple of solid drivers, and they’re getting a lot of practice, so it could go either way.
I like this “drivers vs. programmers” thing, it’s building up a nice rivalry! Interestingly, one of our best drivers is the little brother of our main programmer. I bet it gets ‘interesting’ at that dinner table in a few weeks…
Well, if you really are faster, then record all the settings while you drive, and then replay that as your auto. Seems valid to me.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.