roboteam_suite

This repo contains:

• Roboteam AI, which makes decisions on what the robots should do based on the vision data, and feedback data (received from robots) which it receives from Roboteam World and Roboteam Robothub respectively.
• Roboteam World, which interprets vision data (received from GRSim, or from a physical camera)
• Roboteam Robothub, used to send commands to the robots (either through GRSim or the base station)
• Roboteam Utils, which contains a lot of helper functions, primarily geometrical in nature, such as Lines, Triangles, etc. These are used in other repo's, such as Roboteam AI and Roboteam World
• Roboteam Proto, which contains the protocol used to send and receive messages over TCP/UDP

Please note: This software is tested on Ubuntu 20.04 LTS, and the installation instructions are written with that distro/OS in mind. However, other distro's of Linux are also likely to be able to run this software, albeit with some modifications.

Usage

First, clone this repository. We recommend using an SSH key if you are able to do so, as this immediately authorizes your future pushes to the repositories.

$ mkdir -p ~/roboteam
$ git clone --recursive https://github.com/RoboTeamTwente/roboteam_suite.git ~/roboteam/roboteam_suite

Now, you can open the roboteam_suite project with an IDE of your choice. We recommed CLion, as it has a lot of very helpful features and you can get a student license for it.

Make sure you have the dependencies installed. Go to the CMakeLists.txt and click on it. CLion will generate a pop-up saying "project files not loaded" in blue in the top right corner. Clicking it will have CMake load the project. This means it generates instructions on how the files should be compiled in order to produce the executables specified in the CMakeLists.txt file. You can now compile the targets!

The easiest way to run the application is to make some compound executables in CLion that boot everything you need. To run 1 team in GRSim, the simulator, you will need to add roboteam_ai, roboteam_world, and roboteam_robothub. Make sure you have GRSim or ssl-vision running and are listening to the right ports. In general, the GRSim vision multicast port should be set to 10006.

Auxiliary Software

It is recommended to use the game controller to control the gamestates. You can do so by downloading the latest binary release of the SSL-Game-Controller repo, and executing it. For example, when downloaded to the Downloads folder:

~/Downloads/ssl-game-controller_v1.2.0_linux_amd64

You might need to chmod the file to make it executable. The AI client should now listen to the commands sent by the GameController.

Also, make sure you have installed GRSim using the instructions found on the RoboTeam fork of GRSim. Now execute the grsim binary from the command line. This usually looks like this:

/directory_you_cloned_grsim/grSim/bin/grsim

Making the executables

Make sure the vision multicast port is set to 10006

One Team

Make the compound executable shown below: Then run this compound executable and run GRSim.

Two Teams

Follow the steps used for the other compound executable, only now add 2 extra targets: roboteam_ai_1 and roboteam_robothub_1, as seen in the screenshot, to the compound. While creating the new CMake Targets, be sure to include a "1" in the program arguments, as seen in the image.

Installation Ubuntu

List of dependencies

• CMake >3.16
• Protobuf 3.9.1
• ZeroMQ 4.2.5
• Eigen3 3.3.7-2
• QT5
• QT5Charts
• Ninja
• CCache
• Google Test and Google Mock

Either install these manually or run the script called install_UBUNTU_20_04.sh

Installation on macOS (tested on macOS 12.2.1 Monterey)

Make sure you already have the following:

• XCode
• XCode Command Line Tools (Run XCode once, then xcode-select --install)
• Homebrew

Install remaining dependencies

$ brew install cmake zmq armadillo libsodium qt@5 protobuf llvm

Create a sparse (bundle) disk image with a case-sensitive FS to hold the source code

$ hdiutil create -size 2g -volname rbtt-workspace -layout GPTSPUD -fs "Case-sensitive APFS" -type SPARSEBUNDLE -nospotlight ~/Desktop/rbtt-workspace
$ hdiutil attach ~/Desktop/rbtt-workspace
$ cd /Volumes/rbtt_workspace/
$ git clone [email protected]:RoboTeamTwente/roboteam_suite.git
$ cd roboteam_suite/
$ git submodule update --init --recursive
$ git submodule foreach git checkout development

Select a newer clang to be used to compile RBTT software

1. Open CLion
2. Go to Preferences -> Build, Execution, Deployment -> Toolchains -> Default
3. Change C Compiler to: /usr/local/Cellar/llvm/13.0.1/bin/clang
4. Change C++ Compiler to: /usr/local/Cellar/llvm/13.0.1/bin/clang++

Notes:

• C(++) Compiler paths may differ based on installed version. Verify the version with $ brew info llvm
• Toolchains may vary if you changed the default CLion settings. If you want to deviate from this readme, ensure the compiler selected in the toolchain used to compile RBTT software has support for C++20

Code Style Formatting

To be consistent in the code style, we use clang-format. You can use clang-format to automatically format your code during every commit and you can use clang-format in CLion to format all code files in a given folder.

You can install the pre-commit hooks by doing the following

$ cd ~/roboteam/roboteam_suite
$ ./formatting/install.sh
# Should give output:
clang-format has been properly set up

CLion Formatting

• Open CLion and go to File->Settings->Tools->External Tools
• Click on Add and fill in without quotes:
  • Name: 'clang-format'
  • Program: The location on your computer to the formatting/format-all.sh file of this repository.
  • Working directory: '$FilePath$'
• Click on OK, OK
• If you now right click a folder in your Project Tree and go to External Tools then you can click on clang-format which will format the entire folder (not this unfortunately does not yet work on single files).

See also

VSCode

To run roboteam code within VSCode, which John thinks is nicer, but nobody else agrees with :(, you can always ask him for help on setting up.

Running with camera's and real robots:

• SSL-Vision

Working with a referee

• SSL-Game-controller
• SSL-vision-client

Working with the grSim simulator

• grSim

Tracy Profiler

Tracy is quite cool(and lightweight) profiler, that can help you analyze the performance of your code.

To enable Tracy

• Define TRACY_ENABLE macro (e.g. by passing -DTRACY_ENABLE=1 to cmake)
• Add ZoneScoped/ZoneScopedN etc. to scope you want to evaluate
• Compile Tracy Server and run
  • Information is in the tracy docs
• Run AI