Reliable Autonomous At Worlds

Each year we go to worlds we face issues with our autonomous due to our journey. Is there any way of doing autonomous is easy to implement whilst also being reliable.

2 Likes

Not speaking from experience, but most of the time I’ve been lazy with auton despite knowing how to make it as reliable as possible.

  1. Use sensors on everything. Time should almost never be considered for auton other than moving on to the next step and inside PID.
  2. PID helps, which helps with repetition.
  3. Make sure you always run auton with a charged battery.
  4. Our autonomous became super inconsistent when we utilized light receiving sensors such as the light sensor, line sensor, and vision sensor. Make sure to run auton in a manner that it is consistent, even in a 100% pitch black environment because lighting is different in every field.
  5. Repetition. Repeat your auton over and over again, 5-6 times. Record what they did every time and change one variable at a time.
  6. Tracking wheels are really hepful.
  7. If you are using PROS, the okapi contains integrated odometry. Theres a 5225A document showing how odometry works and how to utilize it.
  8. Never make any major changes after your auton is made.
8 Likes

When you get to worlds make sure all your sensors work. Have code that can verify a sensor has malfunctioned.

During ITZ one of the optical shaft encoder broke during shipping. Auton was a mess. The team didn’t think to check if sensors were reporting sane values. The coder spend the better part of a day just to get a reliable cone on the high goal. I felt really bad for him since he spend maybe 40 hours the two weeks before ensuring consistency of MB + 2 cones in 20,10, or 5 zone based on partner or side.

5 Likes

If you hadn’t shared that story, I was about to. I remember you telling me that. Make sure to keep your encoders safe.

1 Like

Redundant systems.

  • Optical shaft encoders should match the inertial sensor should match the motor.
  • Line readers that you can calibrate on the field when preparing for a match.
  • Well selected timeouts.

Also, you need to know how and why your autonomous fails. Run it over and over in the following situations to see how it can go wrong:

  • Rotate your field 90-180 degrees.
  • Low battery
  • Overheated/overworked motors
  • Slightly spread out blocks
  • Slightly misaligned blocks
  • Slightly crooked block(s)
  • Tiles flipped upside down or coated in antistatic spray (varying the rolling friction)
  • Field perimeter moved 1 inch in any direction
  • Some of your deploy rubber bands broken
  • Friction on one side of your robot (just put a weight on it to simulate the problem)
  • Purple neutral block pushed onto your side w/ force of another robot behind it
  • The wrong auton selected (to make sure you do not mess with your teammate or commit a foul
  • Shaft coming out of one of your drive motors (if that has ever happened to you)
  • Bad luck (or any other things you can think of)

If you can prepare for every way you think things could go wrong, you will have a good auton. Not only will it be programmed to handle issues, but you will have seen the ways it can fail, and can quickly fix or address problems between matches (or at the practice field)

8 Likes

In terms of the implementation of odometry, is it hard for the average programmer to implement it effectively. How long will it take to implement ?

Depends if you want your robot to know exactly where it is, or if you just want some redundant sensors to verify that your robot did what it set out to do.

Full odometry take a good bit of math understanding just to approach. However with no other robots messing with you during autonomous, good vision sensors might be a better approach. If your robot always lines up its intake to the blocks, then you don’t need to worry about where it is. Just make sure it has optimized tilt and drive velocity and you should be good.

1 Like

Not as much of a problem this year with V5, but make sure your motors are fine too. During TP I had a solid 7 point auton and it worked great… until states when our motors started to fail miserably. Slowly our 7 point auton trickled down to a 0. Also, wheel stripping is a killer, I had one on my wheels get stripped(we now have the metal insert on) and when you tried driving straight, you got a very noticeable curve.

1 Like