Skip to content

Latest commit

 

History

History
55 lines (46 loc) · 1.65 KB

LESSONS.md

File metadata and controls

55 lines (46 loc) · 1.65 KB

Lessons Learned

Lifter system

Discussion topics

  • why dual pid? - current limiting on controller, and positional control with simple PI controller
  • where's our intake sensor?
  • homing process / auto-correcting belt skips
  • ideally move encoder to directly read output shaft
    • belt skip doesn't matter
    • because it's clearly better
    • Less sensitive to gearbox changes
  • belt feed and forward intake autostop
    • why we didn't use it
    • why it worked awesome

Shooter systems

Discussion Topis

  • Linking the Shooter with the Intake
  • Shooting Prossess
  • Feed Forward
    • Why it worked well
    • Why it didn't
  • Motor Slaving
  • PID
  • Encoder
  • Disabling forward velocity on PID to improve efficiency

Chassis

Discussion topics

  • PID was awesome, kinda
    • need time to tune properly
    • needs filtered output; Oscillations burn out motors, gearboxes
    • feed forward is necessary for responsiveness
    • worked really well for drivers and consistent driving
    • Those encoders are just, the worst.
    • Gyro was cool, and helpful when it worked
      • issues with turning the correct direction
    • brownouts are awful, and worse when PID control/gearing is capable of causing it
      • auto-shifting + current detection kinda helps

##Vision

Discussion topics

  • Everything is awful!
    • static IPs are absolutely required on everything
  • Brownouts on Raspi cause problems
  • field lighting is the worst.
  • ideally, do more advanced detection of retroreflective based on flashing LEDs and comparing frames
  • Make sure tools can rotate camera without incurring delays and problems BEFORE creating fixture to rotate camera
  • always update the tools

Robot Behaviours