Get The Maximum Out Of The Sandstorm Period


#1

Hey, I’m curious what are the talks in your team regarding the new autonomous period.
Are you leaning towards letting your drivers control the robot during autonomous period or operate the robot autonomously?
I’m thinking robots driven by a driver might get in the way of a robot that drive autonomously, plus there’s that ramp that messes up encoder count…
But … with autonomous code we can get faster more accurate results.
We also came up with an idea of maybe driving to the ship manually, placing a panel autonomously, and then driving to the loading station and back to the ship to place a panel autonomously, the driver might even manually tell the robot to which bay place the second panel.
I was wondering what are your plans for the sandstorm period?


#2

We’ve never been great at autonomous so we’re planning on just driving.


#3

A really cool idea my team had was to do everything teleop style, but to have a button called “auto” on our control station and essentially program autonomous routines to that button, but to cancel out in the event the driver inputs another control, just in case something goes wrong.


#4

That’s a good point to think about being a teleop bot in the sandstorm period could cause some issues, especially when having all autonomous robots have collisions pretty often. Having an autonomous option would make sense to strategize with alliance partners to make sure you won’t collide even more. I can’t imagine teams would be that big of a fan if a driver hit them when they’re autonomous.


#5

Coming from a team with historically good autos, this is the strategy we are most likely to use, as it minimizes margin for error. (Plus it allows us to use the auto bits to assit in tele)


#6

We’re currently leaning towards writing autonomous routines that as previously stated, have an override button for the driver so if something goes wrong or if mid-sandstorm we wanted to do something else, we have the opportunity. The way we view it is that the majority of the time, an autonomous routine can complete an objective faster and more repeatably than a driver can.


#7

I expect most teams to do teleop with cameras and maybe other sensors, especially as most teams have run auto routines at a quarter speed or less, and been done with that five seconds in; teleop will let them do a lot more in that fifteen seconds. My first blush strategy would be to start with a hatch cover on the robot, and after placing that, start moving cargo from the depot into the cargo ship, where it won’t roll out, and add hatch covers after the sandstorm passes.

I fully agree - but those teams capable of pulling off high speed, multiple objective autos aren’t nearly as common as those who can get a driver camera streaming.

Edit:

Oh, misread that. I thought that was the end of the match. Yes, at t-135s, not t-0s.


#8

This is something everyone is overlooking. Think of the 2, 3, and 4 cube autos that happened last year. Even though they didn’t happen all the time, they probably successfully occurred more often than a human controlling the robot in the same time frame.

My prediction is that the lack of a required auto mode will not stop the programming capable teams from making Autonomous routines that complete objectives faster than humans can.


#9

Thanks for helping out guys.
I really wanna get this part of the game right so my team won’t burn precious time…|
I’m still wondering though as to how planning of the sandstorm period will go, do you think maybe an alliance can plan out this period so they won’ t collide or do you think it’s to hard to plan the sandstorm period in advanced?
Plus, obviously a good sandstorm routine will give your alliance some sort of lead, but do you think that this year, the sandstorm period will be as significant as the autonomous mode the years before?


#10

Any cargo not secured by hatch covers at the end of the sandstorm will roll out.


#11

Oh, absolutely. What my statement was suggesting is that teams should consider still automating some amount of their process if they have the capacity to. Though there is no required auto, time teams spent on autonomous modes in past years should be spent on automating individual processes. In my opinion, the programming teams should be just as busy this year, if not more.


#12

Is it known that we will have the ability to override the autonomous routine if needed during the sandstorm period?


#13

It depends on how you program the auto. The period accepts driver input, so you could conceivably have something in your auto code that checks for that input and stops execution/switches to a different routine.


#14

Yes, you can actually switch completely to teleop in auto if programmed correctly. You can also alter the auto program using a joystick if you program it that way as well.

If you can think it, you can probably do it with programming. Some things are harder than others to do.


#15

It will be interesting to see how driving off tier 2 (maybe ever tier 1) will effect full auto and pathfinder routes.


#16

Our software team is pretty adamant that fully autonomous is the way to go. I have concerns with getting off the platform and staying aligned, but if worse comes to worst we can just do that part teleoperated.


#17

You can always have your software team set up a gyro and have the robot reorient itself once it gets off the platform. Chances are, your Roborio already has a gyro on it.


#18

You can use navx to create a straight drive function. We did this in 2016 for crossing the defenses and were able to create a pick line hitting the robot in different positions and it kept correcting to it’s original path.


#19

While I’m not sold on the teleop section of the Sandstorm, what I do like about it is it really added in the “fail safe” that a lot of people were asking for during the offseason. If your auto starts to fail, you now have the option to bail out of it gracefully instead of watching hopelessly as your robot destroys itself.


#20

Keep in mind that pretty much every hatch cover and cargo load is going to be in “sandstorm” mode, even when you’re in teleoperated mode. The view from the driver station will make it very difficult to align your robot in teleop mode without using some form of on-board video assistance.

The rewards for having efficient hatch placement/cargo scoring in sandstorm will play out for the entire match. I would go so far as to suggest that a team with limited resources would be wise to choose between hatches or cargo in order to finish their robot early and have at least two weeks to practice and optimize their cycle times and reliability for scoring their chosen game piece rather than try to do both and miss out on practice time. Give your coders and drivers a chance! :slight_smile: Doing one thing well is often far more useful than doing two things not-so-well.

Jason