Design
Our robot for the DECODE season, Bumblebee (ARCHEOBOT XYZ), is the fourth iteration in a season of constant refinement — each build measurably better than the last.
Drivetrain. We moved beyond our usual configuration to a belt-driven drivetrain: two 312 RPM goBILDA motors in the back drive 104 mm GripForce mecanum wheels up front. After discovering at the State Championship that heavier motors left us underpowered, we stripped the robot down relentlessly — the final drivetrain weighs just 4.1 pounds total, which keeps us fast and lets us take full advantage of Pedro Pathing’s predictive braking. Our side plates evolved across generations from PLA to aluminum to a custom stainless steel drivetrain for a stronger, more defensive robot.
Turret & launcher. A rotating turret powered by a 6000 RPM motor lets us shoot from different angles and distances. The turret carries a hood that automatically adjusts the flight path by distance, eight goBILDA steel flywheels that hold inertia for rapid-fire shooting, and a gate that shortens the path between artifact and flywheel. We settled on a 30–50° launch angle to cut backspin and shoot accurately.
Intake. Across three generations our intake went from heavy and narrow to a lightweight design using surgical tubing on the ends with gecko wheels in the middle. Funnels with a sharp-then-shallow angle pull artifacts out of corners and line up three at a time.
Endgame & durability. A kickstand raises the part of the robot that sits outside the parking zone, maximizing our endgame score. Absorbent bumpers protect the chassis and odometry after a hard collision dented an earlier build.
Strategy
Autonomous. We score our 3-artifact preload, then the 9 artifacts on the field plus more from the gate — 18+ artifacts — before parking, all inside the 30-second window.
TeleOp. Depending on our alliance we run one of two plans: collect from across the field and the human player and strafe to shoot from the far zone, or cycle quickly through the close launching zone. In both, we use our size and strength to play controlled defense. We favor rapid-fire shooting over spindexers, which we’ve found slower and less reliable to maintain during qualifiers.
Endgame. In the final 20 seconds we keep scoring, then in the last 5 seconds park half the robot inside the zone and use the kickstand for the rest.
Programming
We program in Java in Android Studio, organized into reusable subsystems (turret, intake/transfer, drivetrain) shared between our autonomous and TeleOp — the standard architecture used across FTC. Autonomous paths run on Pedro Pathing, an open-source pathing library whose predictive braking lets the robot slow down and travel faster.
Our turret aiming math computes the distance to the goal, predicts where an artifact will land while the robot is moving, solves for the turret angle, and looks up the right hood angle from an interpolation table — letting us shoot accurately on the move.
Thanks to mentor Danny Diaz and our experience with the current Control Hub, we were selected from thousands of teams worldwide to alpha/beta test FIRST’s next- generation control system — the SystemCore and MotionCore, along with the A301 actuator. As we build expertise with the new system, we’re sharing what we learn across the FIRST community.