This is a study into creating a fully automatic FRC robot.

This process relies on several key algorithms:

  • pose estiation: using the WpiLib DifferentialDrivePoseEstimatorClass
  • global position updates: I am using an algorithm called UcoSLAM that takes a camera feed and outputs position in 3d space
  • pathing planning: Informed Random Rapidly Expanding Tree *
  • Traveling Saleman searches to find the fastest way to collect multiple games objects
  • HSV thresholding to detected yellow balls

Future potential features:

  • robot detection (either ML or detect color of bumpers)
  • intelligent defences
  • avoiding moving obstancles while path planning
    • kalman filters


View Github