Artifacts stuck in the center
With a single intake, artifacts got stuck between the shooters. We built a wall between them, programmed a way to push stuck artifacts out, then designed a diverter to guide them.
Past season · 2025–26
DECODE robot
Built on a goBILDA base, then iterated into something all our own: custom shooter ramps, a gecko-wheel intake, and camera-guided aiming.
Gameplay strategy
How we planned to win
Right after kickoff, we built our whole season strategy around the ranking-point rules.
#1
Movement
A developed, scoring autonomous, plus parking in endgame.
#2
Goal
Subsystems that can intake and shoot quickly, and aim.
#3
Pattern
Program and strategize for a high-functioning motif shooter.
Intake & shooting
From field to goal in four steps
- Step one
Gecko wheels pull the ball inside the robot, guided by where the diverter is set.
- Step two
Balls are held in place by bumper wheels and small boot wheels.
- Step three
A second set of larger boot wheels kicks the balls up into the shooting zone.
- Step four
1:1 6000 RPM motors launch the ball up our custom ramps and out to the goal, all positioned by the camera.
Intake
Problems we solved
Subsystems out of sync
Our subsystems didn't work together easily because of voltage and the number of steps. Drive practice plus better controls meshed every step together.
Intake malfunctions
The first stage's gears kept shifting off each other. We designed a separation brace to support them so they couldn't shift.
Output
Five iterations of our shooter ramp
We traced our prototypes, found the arcs they were made of, and used geometry and trigonometry to design the exact lengths and angles.
Iteration 1
- Created from the goBILDA RITD CAD
- 1.5 mm polycarbonate and zip ties
- No math yet; needed a thicker, more accurate design
Iteration 2
- Tested polycarbonate and cardboard prototypes
- Traced them and collected data
- Correct angle, but needed more support
Iteration 3
- Second design was too low for the goal
- Repeated the math with PLA in mind
- Worked toward a permanent design
Iterations 4 & 5
- Less flex while shooting
- Cut-off bottom for a smooth hand-off
- Thinner for less friction; screw supports added
Software
Code that makes driving easy
One-button aiming
We condensed our controls to three buttons per controller. One button lines up the robot, sets the right flywheel velocity for the distance, and spins up the flywheels, all at once.
AprilTag camera
The camera uses bearing to AprilTags to position the robot in auto and teleop. Aiming time dropped a lot and our shot ratio improved.
LED indicators
goBILDA prism LEDs turn red when the flywheels are up to speed, so drivers never shoot too early. Shooting got faster and more accurate.
Far autonomous
- Shoots 3 preloaded artifacts
- Collects 3 artifacts from the human player zone
- Shoots 3 artifacts
- Parks outside the start zone
Close autonomous
- Backs up to shoot 3 preloaded artifacts
- Collects 3 artifacts from the sorted line
- Shoots the artifacts
- Backs up to park outside the launch zone
We use the IMU to position ourselves on the field before switching to the camera, so small setup errors don't affect the auto. At our summer program, Maddie found that drive-by-gyro beat encoders for accuracy.
Engineering design process
How we design
Define & analyze the problem
Brainstorm initial ideas
Research
Design & prototype
Collect & analyze data
Iterate & improve
Data collection & analysis
How we know our priorities
We use Failure Mode and Effects Analysis (FMEA) all season to decide which projects to prioritize, on and off the robot. Here's one sample from after our first qualifier.
| Requirement | Failure | Cause | Occur | Severity | Detection | RPN |
|---|---|---|---|---|---|---|
| Speed of process | Dead space | Design too spaced out | 7 | 6 | 6 | 252 |
| Creative | No CAD | Not iterating enough | 5 | 7 | 5 | 175 |
| Consistent velocity | Motors not having time to speed up | Strategy / drive practice / camera / program | 5 | 5 | 5 | 125 |
| Creative | Robot in three days | Not iterating enough | 4 | 5 | 5 | 100 |
| Process different | Diffuser inconsistent | Program / servo | 9 | 4 | 2 | 72 |
Autonomous trials & shot tracking
During testing we log every autonomous trial: artifacts scored, crossing the shoot line, collecting the next set, and whether the robot sees each AprilTag. After every third trial, we make a design change.
We also filmed our shooter after qualifier 1 and marked up every shot, so we could see our accuracy in one place before making design decisions.
Professional help: Aaron Bengtson, a biomedical engineer, taught us different types of data collection and their purposes, and helped us customize our process.
Before State
Maximizing robot efficiency
- Tuning and simplifying controls to fit our drivers' skills
- Clear pathways for artifacts through the system, with less dead space for jams
- An extra bar on the robot's rear to open the gate easily
- Timed runs to train our drivers' strategy through the match
- Boot wheels that spin backward automatically so artifacts can't be shot before we're ready
