BasicSim

A minimal ROS2 robotic simulator, with emphasis on minimal. Meant to replace Stage for applications where you only need extremely basic 2D movement (since Stage in not supported in ROS2).

Does:

• Publishes /clock, with the ability to speed up or slow down time.
• Publishes a static map -> odom TF for when you don't need to simulate actual localization.
• Publishes odom -> (robot)_base_link TF for each robot.
• Publishes /(robot)/odom topic with ground truth pose and velocity.
• Listens to /(robot)/cmd_vel and applies the velocity commands to move the robot.
• Reads ros2 OccupancyGrid files (same format as map_server) and keeps the robot outside of walls.
• Keeps track of multiple robots in the same simulation session.
• Simulates 2D laser scanners and publishes to /(robot)/(laser_name). You can have multiple ones on the same robot!

Does not do:

• Physics.
• Velocity or acceleration limits.
• Any sensors other than 2D laser scanners.
• Robot shapes (always assumed to be a circle of given radius, defaulting to 0).
• Robot-robot collision.

Installation

Simply clone inside or your colcon workspace's source directory and run colcon build --symlink-install.

⚠️ This repo has sub-modules! Remember to clone with the --recurse-submodules option, or to run git submodule update --init --recursive.

Configuring a simulation

The entire configuration consist on creating a .yaml file. For example:

map: "occupancy.yaml"
robots:
  - name: "giraff1"
    position: [2, 3, 0]
    angle: 1
    radius: 0.25
    sensors:
      - type: "laser"
        name: "laser_scanner" 
        minAngleRad: -1.9
        maxAngleRad: 1.9
        angleResolutionRad: 0.01 
        minDistance: 0.1
        maxDistance: 4.0

  - name: "giraff2"
    position: [3, 6, 0]
    angle: 2

The map field holds the path to a ROS-format occupancy map yaml file (such as the ones generated by map_server).

robots is a list, where each entry defines a separate robotic agent. Inside of each robot's entry, you must specify a name and starting pose. Optionally, you can also attach to each robot a list of sensors. Currently, the only supported sensor type is 2D laser scanners, but who knows what the future might hold.

Node parameters

Once the simulation is configured with a yaml file, you can launch the simulator with the following parameters:

basic_sim = Node(
            package="basic_sim",
            executable="basic_sim",
            parameters=[
                {"deltaTime": 0.1},
                {"speed": 5.0},
                {"worldFile": "[path to file]/simulation.yaml"}
            ],
        )

• deltaTime is the length of each simulation step (in simulation time). That is, for deltaTime = 0.1s, if a robot is told to move at 1m/s, it will move 0.1m in one simulation step, independently of how long that simulation step takes to run in real time.

• speed is how fast the simulator tries to run, compared to real time. speed = 1 means one simulation step happens every deltaTime seconds. speed = 10 means it will run 10 times faster than real-time. Obviously, at some point processing power becomes a limiting factor (although basicSim is pretty lightweight).

• worldfile is the path to the yaml file we looked at in the previous section.