What impressive things did you do in software this year?

Now that the season is over, our team still has a bunch of software kids that want to get better and continue developing things. I am looking for ideas of particular software elements that we could improve, and am curious to hear what other teams implemented this year.

I have heard in the past:
-Autonomous Driving in Teleop
-Alternate Control systems to PID

5 Likes

We want to figure out motion profiling/trajectory following. I suggest you do too.
What level are you at right now?

3 Likes

Ball tracking via computer vision and realtime system-wide current control were my favorite software features of our 2020 robot, Sir Ballin.

2 Likes

Custom visualizer to generate quintic splines and velocity profiles for motion profiling in autonomous

5 Likes

My favorite was using Logic Diagrams to figure out when we wanted to run our Pickup, Intake, and Launcher subsystems based on Joystick Button inputs and Sensor Inputs and translating that into using the Command-Based Trigger class’s and(), or(), and negate() methods to schedule commands rather than putting the logic in if statements in other commands.

3 Likes

One of the first things we did with our limelight when we got it in the summer…

1 Like

Please explain or post code… It sounds interesting and useful for us, but I want to see what exactly you mean…

2 Likes

Capture replay autonomus, where you record the auto via driving, and play it back as the autonomus.

8 Likes

Our code is always Open Source.

Around lines 120-140:

5 Likes

We wrote some pretty darn accurate vision code.
.
.
.
.
.
.
.
.
.
And we didn’t use a limelight

27 Likes

Thanks, seems interesting. Will look into it for next year/offseason/resumed competitions.

Posted this in our robot reveal thread:

4 Likes

We built a statistical model based on experimental data to always hit the inner goal (at least with the balls we had). The robot would get the vision target, aim the hood, and align the entire robot with the target.

Edit: Chameleon Vision for the win!

2 Likes

We had motion profiling and a limelight working well enough that we would have consistently scored 4-6 cell autos and scored very well in teleop. It was our best bot by far. No bag day has been an absolute blessing for us. It sucks to be graduating when you know your season would have been good.

2 Likes

Current control, as in limiting current on motors? How did you do that? On Talons?

Actively measuring current draw and allocating current limits on the fly to prevent brownout. Most obvious example is preventing the compressor from running while the flywheel is spinning.

4 Likes

We took the first steps towards getting path planning working. We got a limelight and the students programmed it and set it up (although we didn’t really have a use) all by themselves which is always fun to see

We also took the first steps towards getting path planning working

1 Like

I’m really glad someone actually used those features :wink:

1 Like

It worked a treat, and made perfect sense with the continuing trend of Command-Based for all your top level “logic” being in RobotContainer. Also gave me an excuse to teach some basic boolean algebra and logic diagrams to the students.

I pretty proud of auto this year, we used a few different techniques. We used a lime light for getting/tracking balls and for shooting. Encoders to get from point A to point B and a gyro to make our turns more accurate.

1 Like