Skip to content

Latest commit

 

History

History
108 lines (71 loc) · 4.37 KB

README.md

File metadata and controls

108 lines (71 loc) · 4.37 KB

til-final

SDK, simulator and documentation for TIL 2023 Robotics Challenge.

  • src/: SDK and simulator packages. YOU DO NOT NEED TO MODIFY THESE.
  • config/: sample configuration files for the scoring server, simulator and sample autonomy code. You might want to copy these.
  • data/: sample audio and images for the AI tasks, sample maps used for. the path planner and simulator. Feel free to use these data samples.
  • docs/: documentation and Sphinx docs source. Build these with the instructions below to view API docs.
  • stubs/: Code stubs for participants to implement. It is recommended to develop your custom codes in this directory.

Setup

Step 1: Clone this git repo.

It is highly recommended to install python dependencies in a virtual environment.

Step 2: Activate virtual environment.

pip install virtualenv  # you do not need to run this line if you are down at our physical venue.
virtualenv -p python3.8.10 venv  # create virtual python environment with specific python version.
source <NAME_OF_VENV>/bin/activate  # on Windows, "./<NAME_OF_VENV>/Scripts/activate" to activate virtual python environment.

Step 3: Install the RoboMaster SDK

pip install robomaster

Note that the required python version for the RoboMaster SDK is between 3.6.6 and 3.8.10

Step 4: Install the custom challenge SDK

pip install -r requirements.txt  # install third-party dependencies. use the provided requirements-win.txt if installing on Windows.
pip install .  # install the SDK from source. The '.' means install from current directory.

Step 5: Install a python GUI backend.

On Ubuntu, try this GUI backend:

sudo apt-get install python3-tk  # If you are at our venue you don't need to do this (it is already installed).

or if you're on Windows

pip install pyqt5

Step 6: (Optional) Install any other dependencies your custom AI models require.

For example, the sample AI models require these.

pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu116
export PYTHONPATH='/your/path/to/til-23-finals/stubs/yolov7'  # on Windows, replace 'export' with 'set'

Note: set command sets an environment variable temporarily only. Use 'setx' to set a variable permanently.

Build the documentation

Much of the technical information needed to understand and run the robotics system is provided in the Sphinx-generated docs. Be sure to generate and view them!

# install sphinx dependencies
pip install sphinx sphinx-autoapi sphinx-rtd-theme

# Generate html docs based on source folder.
sphinx-build -b html docs/source docs/build

Access the html docs at docs/build/index.html.

Quickstart Example with Simulator

Open a terminal, start the scoring server:

til-scoring config/scoring_cfg.yml -o teamName1

In another terminal, start the simulator:

til-simulator -c config/sim_cfg.yml

In another terminal, start your autonomy code:

python stubs/autonomy_starter.py --config config/autonomy_cfg.yml

Note: REMEMBER to CHANGE THE FILE/DIRECTORY PATHS in the config files (by default in the "config" folder) to your own directories. And to modify the config files as necessary.

You are free not to use any of the provided stubs or code, except for the tilsdk's ReportingService and LocalizationService, which are mandatory for you to interact with our challenge servers.

You can always use the --help option with these commands to view help messages. E.g.

til-scoring --help

Start developing

To prevent your code from being overwritten by patches and code releases by the organizers, you should make your own copies of the config files in the config/ directory. and also your own copy of the stubs/autonomy_starter.py if you intend on using the sample code.

Tips / Common Issues

Make sure your source code points to all the correct file and directory paths.

Ensure you are pointing your LocalizationService and ReportingService to the CORRECT IP addresses and port numbers where the servers are actually located.

You are supposed to make various improvements to the autonomy code. You can think about improvements to the robot movements, the path-planning, or the usage of the path planner's output waypoints, or the management of noisy localization data, etc etc.