Skip to content

Run It on the Romi

Everything so far ran on a computer pretending to be a robot. Now the same code drives the real one.

🌐 How the Romi is wired

flowchart LR
    LAP[Laptop<br/>your program, in simulation mode<br/>Sim GUI as Driver Station]
    PI[Raspberry Pi on the Romi<br/>Wi-Fi access point, 10.0.0.2]
    BRD[Romi control board<br/>motors, encoders, gyro, LEDs, buttons]
    LAP <-- "Wi-Fi, websocket" --> PI
    PI --- BRD

The program does not get copied to the Romi. It runs on the laptop, exactly like the simulator, and every motor command and sensor reading crosses the Wi-Fi link. That is why build.gradle has HALSIMWS_HOST set to 10.0.0.2: it is the Romi's address on its own network.

Romi Competition robot
Program runs on Laptop roboRIO
Network Romi's own Wi-Fi, robot at 10.0.0.2 Robot radio, roboRIO at 10.TE.AM.2
Driver Station Sim GUI window FRC Driver Station app
Deploy step Just run it ./gradlew deploy copies the program to the roboRIO

1. Connect the laptop to the Romi

  1. Turn the Romi on. Wait about a minute for the Raspberry Pi to boot.
  2. On the laptop, join the Romi's Wi-Fi network. Its name starts with WPILibPi. A mentor has the password.
  3. Open a browser to http://10.0.0.2.

You will see: the Romi's web page. The Romi tab shows the board is connected. This is also where a mentor calibrates the gyro if it drifts.

You are off the internet

While on the Romi's Wi-Fi the laptop has no internet. Fetch your branch first, then switch networks.

2. Check out your branch

Run these on the laptop, with your branch name, before joining the Romi's network:

git fetch origin
git checkout yourname/session-5

3. Run

Run the robot program:

./gradlew simulateJava

You will see: the Sim GUI opens as usual, and the terminal prints a line like HALSimWS: WebSocket Connected. That is the link to the Romi. If it prints connection errors instead, check the Wi-Fi.

The yellow LED on the Romi is solid. That is the Session 3 RSL rule, in Disabled mode. Challenge solutions never reach main, so if you want to see your SOS on the real red LED, check out your yourname/session-3 branch on this laptop.

4. Drive

  1. Drag the controller onto Joystick[0].
  2. Click Teleoperated in Robot State.

You will see: the yellow LED starts blinking. Push the left stick forward, gently. The Romi drives. Right stick turns it. Hold the right bumper: slower. Hold A: the green LED lights. All of that is Session 4's code, which came to main when the mentor merged it. Did the Session 4 challenge? Check out yourname/session-4 on this laptop, press its button, and watch the Romi turn around. Now is the time to tune its constant, in your Codespace, on that branch.

Put the Romi on the floor or a large table with someone spotting it. It is quick.

5. Check the gyro sign

This is the one thing the simulator could not tell you.

  1. Click Disabled. Open AdvantageScope's field view on Field/Robot, or watch Drive/headingDeg in the NetworkTables panel.
  2. Pick the Romi up and turn it a quarter turn to the left (counterclockwise, seen from above).

You will see: Drive/headingDeg changes by about 90. If the robot on the field view turned left, kGyroReversed = false is correct. If it turned right, the gyro counts the other way: change kGyroReversed to true in your Codespace, commit, and push.

6. Watch real sensors

Drive a straight line and watch Drive/leftInch and Drive/rightInch in AdvantageScope. They should climb together. Spin in place: one goes up, the other goes down, and headingDeg sweeps.

This is the whole testing ladder in one afternoon: the tests said the math was right, the simulator said the wiring was right, and the Romi said the sign was wrong or right. Each level catches what the one below cannot.

✅ Done when

  • Your code drove the Romi.
  • You saw the RSL go from solid to blinking when you enabled.
  • You checked the gyro direction and fixed kGyroReversed if needed.