/Bragi

The repository for Crescendo and our robot Bragi.

Primary LanguageJava

Bragi

Bragi VisionTable feature branch

Robovikes uses the Limelight Coprocessor and PhotonVision for robot computer vision. This repo is for testing and experimentation for simple Limelight tasks.

Limelight Standard Operating Procedure

Here is a standard operating procedure for getting a Limelight using photonvision working for the RoboVikes team.

Background

The Limelight uses its configuration to identify which objects in its image is most likely to be the target. It then sends this information to the RoboRIO where the main code is running.The Limelight is a raspberry pi compute module with a camera

For information about how to install the Limelight software and access the dashboard, visit the official Limelight documentation

Necessary Equipment

Determine Limelight version:

  • Open the limelight by removing the 4x 1.5 mm start screws
  • Remove lid
  • Locate version info. The version should be v1, 2, 2+, or 3.

Robovikes has Limelight v2 and v3

Install software

V2, V2+

  • Wire it up following these instructions https://docs.photonvision.org/en/latest/docs/installation/wiring.html Network according to the diagram below (https://docs.photonvision.org/en/latest/docs/installation/networking.html)

  • Add passive 12 volt passive PoE ethernet to the coprocessor side or power it directly using +12V to the PDV/PDB or or a power bank on the micro USB (if allowed by this year’s rules).

  • Connect the Limelight to a network (through radio, router or switch) that is shared by the laptop.

  • Power on Limelight

  • Connect to RoboRio

  • In browser connect to photon vision at the address photonvision.local:5800

  • Set Team Number to 701

  • Import Settings

    • Hardware Config hardwareConfig.json (always use v2)
    • AprilTag Layout AprilTagLayout

  • Set Static IP to 10.7.1.11 for first camera and 10.7.1.12 for second camera (if needed)

  • On Dashboard

    • Under Cameras, turn off Driver Mode
    • Click APRILTAG
      • Select AprilTag 36h11 (6.5in)
    • In the camera pipeline window, use the interface to update the camera name and pipeline.
      • Note: the camera name must match the name specficed in Constants.java
      • Rename Camera to limelight_v{version number} where verison number can be 2 or 3 (default is Camera_Module_v1 and each camera name must be unique?)
      • Rename Pipeline to AprilTag2024
      • Select Type AprilTag

Calibration

  • Print the PhotonVision chessboard 8x8x1 calibration sheet
  • Measure the square with calipers - should be close to 1". Write actual value (e.g. 0.97") in.
  • Tape the sheet to a clipboard or piece of wood with a large border.
  • Follow calibration instructions.
  • Go to Cameras tab
  • Start calibration.
  • Move board around until colored lines appear and collect an image. Repeat at least 12 times.
  • Save calibration.
  • Repeat calibration for other resolutions. The higher the resolution, the lower frame rate.

Port Forwarding

  • In the Robot.java code, add the following line to the robotInit() function.

Add PortForwarder.add(5800, "photonvision.local", 5800);

Camera Streams (For dashboard?):

Cameras will stream on ports 1181 and 1182 for the first camera and 1183 and 1184 for the second camera.

  • Once connected and ready to connect to the RoboRIO, set a static IP to 10.7.1.11 for the first Limelight and 10.7.1.12 for the second Limelight.

Observing results

  • Update VisionSubsystem.java with any Limelight related variables you want to observe
  • Connect Limelight to RoboRio
  • Connect to RoboRio using driver station
  • Open ShuffleBoard
  • Point robot to AprilTags and watch the ShuffleBoard update

Continuous integration (CI)

CI workflows are run prior to pushing or merging code in github to improve and enforce code standards and quality to downstream users.

  • CI-build.yaml

    • This workflow tries to build the robot code prior to accepting the code in the remote github repository.
    • The build uses the settings.gradle options.
    • If the build fails, then the code is not accepted into the remote github repository.
    • To pass a failing build, fix the code until the build is successful locally, then commit, merge and push your changes.

  • CI-format-checker.yaml

    • This workflow checks the format of the code format for consistency which improves code readability.
    • The build uses spotless settings in the build.gradle file. Among other settings, the googleJavaFormat is used for checking the formatting of java code.
    • Make sure the .gitattributes file is present if code is being developed or committed to the remote branch from different OSs (PC and Mac[Mr. Murphy]).
    • If the build fails, then the code is not accepted into the remote github repository.
    • To pass failing code, run the .\gradlew spotlessApply command fix the code formatting. Run the .\gradlew spotlessCheck command to until the build is successful locally, then commit, merge and push your changes.
    • TL;DR: Run .\gradlew spotlessApply in the terminal in the .\BragiCode directory prior to pushing code to the remote.