Launching Engines in Docker Containers Examples

The folder "examples" contains four examples (listed in the table below) of how to use the DockerLauncher launch command to launch Engine processes in docker containers. In each of these experiments there is a simulation_config_docker.json configuration file which uses DockerLauncher to launch and run one of the engines in the experiment in a docker container.

Example folder




Running a gazebo_grpc engine with a husky model inside a docker container



Running a nest_json engine with a simple nest model inside a docker container



Running a pysim engine with a arm26 model of OpenSim inside a docker container



Running a python_json engine inside a docker container


The images referenced in the table can be manually built from the different dockerfiles available in the folder “/dockerfiles” using docker-compose. For more information on how to use the provided dockerfiles to build custom, modular docker images with the necessary dependencies to run different Engines, see this page.

To run the experiments, users should first, on the host where Engines are to be launched in docker containers, setup docker daemon properly. See this guide for more information on how to do this.

Once this is done, on the host where the experiment is launched, one should run the experiment as usual. For example:

cd examples/<example folder>
NRPCoreSim -c simulation_config_docker.json


The IP addresses of the local and remote hosts have to be taken into account when setting the experiment configuration parameters. For example, in the case of launching a PythonJSONEngine:

  • "DockerDaemonAddress" is the IP address of the remote docker daemon.

  • "ServerAddress" is the IP address of engine running inside of the docker container (in the remote host)

  • "RegistrationServerAddress" is the IP address of NRPCore

Also it must be ensured that ports required to connect to the aforementioned IP addresses are accessible from the respective hosts.