Pylot is an autonomous vehicle platform for developing and testing autonomous vehicle components on top of the CARLA simulator.

Quick start

The easiest way to run Pylot is to use our Docker images. First, please ensure you have nvidia-docker on your machine before you start installing Pylot. In case you do not have nvidia-docker please run ./scripts/

We provide a Docker image to run the CARLA simulator in, and a Docker image with Pylot and ERDOS already setup.

docker pull erdosproject/pylot
docker pull erdosproject/carla

Next, create a Docker network, a CARLA container, and a Pylot container:

docker network create carla-net
nvidia-docker run -itd --name carla_v1 --net carla-net erdosproject/carla /bin/bash
nvidia-docker run -itd --name pylot_v1 -p 20022:22 --net carla-net erdosproject/pylot /bin/bash

Following, start the simulator in the CARLA container:

nvidia-docker exec -i -t carla_v1 /bin/bash
SDL_VIDEODRIVER=offscreen /home/erdos/workspace/CARLA_0.9.6/ -opengl -windowed -carla-server -benchmark -fps=10 -quality-level=Epic

Finally, start Pylot in the container:

nvidia-docker exec -i -t pylot_v1 /bin/bash
cd workspace/pylot/
python3 --flagfile=configs/detection.conf --carla_host=carla_v1

More Information

Getting Involved



pylot.config module

pylot.component_creator module

pylot.utils module

Module contents