I’m curious to see whether other teams think it would be worth it to work on an autonomous program or just stream to driver station and have drivers take control.
How many teams could do 3-4 cubes in 15 seconds in Teleop last year?
2 Likes
How many teams could do 2 cubes in 15s durring auto last year?
1 Like
More than those who could do 2 in 15s in teleop.
1 Like
Design for both.
Run the option that gets you the most points in each match. This might change match to match depending on your alliance partners capabilities.
Having options to choose from it’s never a bad thing. If you never use cameras because your auto is stellar, great! If it’s not running reliably or sensors fail, fall back to drivers and cameras.
If cameras fail, or bandwidth limits make your feed unusable, fall back to auto.
Maximize your scoring potential. Limp through matches when there are failures instead of being dead in the water.
How’s your autonomous programming going so far? If you don’t have time to test it before a regional, plan on driving via camera. But always keep working!
Even then is it “worth it” for most teams to pursue. I know we would have had a tremendously better season least year if sandstorm existed.
So I guess if your in that ~90th percentile of teams, heck go for it.
But if your the rest of us, cameras are looking good.
I know we will be developing some auto code on the side but I have a feeling some vision assistance and good drive practice might go a long way.
We are planning to drive. The extra driver practice will pay off in our view.
If you can actually get a camera to work really well and the driver is comfortable using it for 15 seconds, then driving isn’t a bad option.
However, even if you can’t get that many game pieces scored in that 15 seconds, if you can get one of your autonomouses working and make it robust, then you’re basically guaranteed to get those early points each match.
This is especially important when precision placement is necessary unlike last year. If you’re off even a little bit, the game piece will probably fall off your robot onto the ground.
If you don’t get vision working, using a camera is a great option. But some autonomous code would definitely be helpful because of latency. The camera isn’t going to be as perfect as everyone thinks it will be. One thing great about this year, is you can add both and either only use the camera if autonomous doesn’t work, or even stop your autonomous while it’s running.
I think many teams, if time was spent well in the offseason learning vision basics (and auto basics in general) would benefit more, but its a little late to suggest that i guess.
Auto will reign supreme, to directly answer OP’s question. Many are going to do vision, like JJ, and i have no doubts that it will be successful for them, but the top will still be done with auto routines.
1 Like