What Is This?

Note

This is a collection of lightweight simulated RMF deployments you can run locally on your workstation to test RMF task workflows and deconflictions.

These deployments uses invisibot and fleet_adapter_invisibot in order to skip setting up a physical robot as well as avoid using computationally-heavy Gazebo simulations.

Tip

rmf_gym_lite is best used for the following:

• Help identify potential deadlocks in your RMF Map Building .yaml file
• Allow ease of finetuning based on recommended graph strategies.

Build 🔨

git clone https://github.com/cardboardcode/rmf_gym_lite.git --depth 1 --single-branch && cd rmf_gym_lite

bash scripts/1_pull_docker_compose.bash

Configure 🔧

Follow the instructions below to create a custom scenario:

1. Create a new directory and call it your designated new scenario name.

2. Copy over the following files from other scenarios:

• debug.rviz
• docker-compose.override.yaml

3. Create the following new RMF Building Map File:

• Image file used for layout.
• .building.yaml

Tip

Each charger spawns an invisibot. Set is_charger property to true for waypoints you wish for the robots to start on.

4. Generate nav_graph.yaml for the new RMF Building Map File using the following command below:

bash scripts/traffic_editor/2_generate_nav_graph.bash

Run 🚀

Start without RViz:

bash scripts/2_deploy_docker_compose.bash

Start with RViz2

bash scripts/2_deploy_docker_compose.bash -y

To stop the deployment, run the command below: bash scripts/3_stop_docker_compose.bash

Scenario(s)

triple_H

simple_lift

simple_doors

delivery_queue

lift_queue

narrow_corridor

Maintainer(s)

• cardboardcode