Minimum working example

To simulate a 2-agent scenario:

./example.sh

This will save a plot in gym_collision_avoidance/experiments/results/example so you can visualize the agents’ trajectories.

You can use gym_collision_avoidance/experiments/src/example.py as a starting point to write code for this environment.

../_images/example.gif

Now, you can either read about the Software Architecture or try some of the Use Cases, such as Compare a new policy to the baselines or Train a new RL policy!

Note

The shell script sources the right version of Python and the venv, but if you’ve handled this on your own, you can simply call the python file:

python gym_collision_avoidance/experiments/src/example.py