This is the code that runs our robot for the 2026 FRC season, REBUILT. It's a swerve-drive robot with a turreted shooter that automatically tracks the goal, an extending intake, and a full indexer/feeder path to move game pieces from the floor up to the flywheel.
It's built on WPILib's command-based framework and fully integrated with Team 6328's Advantagekit, so everything is logged and replayable.
- Drive — CTRE swerve (TalonFX drive + steer, CANcoders, a Pigeon 2 gyro). Field-relative teleop, and it feeds the shooter its pose so the turret always knows where the goal is.
- Shooter — a turret that rotates to aim, a hood that adjusts angle, and a flywheel. Distance-to-goal gets fed through interpolated lookup maps to pick the right hood angle and flywheel speed, so the robot can shoot from basically anywhere on the field.
- Intake — an linear extending intake that deploys out, runs rollers to grab pieces, and retracts. Has jam detection and pulse/unjam behavior.
- Feeder + Kicker — the indexer path that moves pieces up to the shooter. Watches motor current to detect jams and automatically reverses to clear them.
- Vision — Limelight on the real robot (PhotonVision in sim) feeding pose estimates back into the drivetrain's pose estimator.
- LEDs — CANdle status lights.
The idea is that the driver mostly just drives, and the robot handles aiming and firing on its own.
By default the turret is always tracking the goal and keeping the flywheel partly spun up. When you hold the right trigger, it locks on for a real shot (and slows the drivetrain down so you can be precise). Once the flywheel is at speed and the turret is actually pointed at the target, the feeder automatically kicks in and starts feeding pieces into the flywheel. Let go and it idles back down.
If the feeder jams, it notices (via current draw), reverses to spit the jam back out, and then resumes feeding — no driver input needed.
| Input | Action |
|---|---|
| Left stick / Right stick | Field-relative drive + rotate |
| Right trigger | Lock onto goal for a shot (slows drive, auto-fires when ready) |
| Left trigger | Extend intake + run rollers |
| Left bumper | Retract intake |
| Right bumper | Slow-retract / pulse intake |
| A (hold) | Lock heading to 0° |
| B (hold) | Force-reverse the indexer |
| X (hold) | Extend + unjam everything |
| Y | Reset gyro heading to 0° |
| D-pad Up | Emergency zero the turret + intake |
Autos are built in PathPlanner and live in src/main/deploy/pathplanner. Paths get stitched together with named commands like LockOnTarget5s, ExtendIntake, and PulseIntake, so an auto is mostly "drive here, pick up, lock on and shoot, drive there."
src/main/java/frc/robot/
├── RobotContainer.java # wires up subsystems, autos, and button bindings
├── Constants.java # ports, tuning values, shooter lookup maps
├── subsystems/
│ ├── drive/ # swerve modules, gyro, odometry
│ ├── Shooter/ # turret + hood + flywheel
│ ├── Intake/ # extending intake
│ ├── Feeder/ + Kicker/ # the indexer path
│ ├── vision/ # Limelight / PhotonVision
│ └── LEDs/
└── commands/ # the actual behaviors, grouped by subsystem
Every subsystem follows AdvantageKit's IO pattern: there's an IO interface, a real TalonFX implementation, and a Sim implementation. Constants.currentMode switches between REAL, SIM, and REPLAY, which is how the same code runs on the robot, in simulation, and back through recorded logs.