KPI Rover ROS2 Software

• KPI Rover ROS2 Software
  • Get Source Code
  • Build and Launch Simulation Using Docker
  • Build and Launch Simulation Without Docker
  • Build and Launch on RPI
  • Other Notes
    • Start Keyboard Control
    • Launch CSPC Lidar Demo
    • Start DDS Server
    • Launch in Docker when DDS is used
  • TODO List

Get Source Code

Clone Repository

git clone git@github.com:KPI-Rover/ros_ws.git
cd ros_ws

Install Packages Using vcstool

vcs import src < kpi-rover.repos

❗ Important: Run all next commands from the root of the ROS workspace (ros_ws folder) unless specified otherwise.

Build and Launch Simulation Using Docker

❗ Important: All pull requests must include source code that has been built and executed using Docker. CI integration following this approach will be added soon.

Build Docker

docker build -t kpi-rover . -f docker/Dockerfile.sim

Build Project

docker run --rm -it \
  --user $(id -u):$(id -g) \
  -v $(pwd):/workspace \
  -w /workspace \
  kpi-rover colcon build

Launch Simulation

docker run --rm -it \
  --init \
  --network=host \
  --user $(id -u):$(id -g) \
  -v $(pwd):/workspace \
  -e DISPLAY=$DISPLAY \
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -v ~/.gz:/home/ubuntu/.gz \
  -w /workspace \
  -e ROS_DOMAIN_ID=1 \
  -e GZ_PARTITION=1 \
  kpi-rover \
  bash -c "
    source /opt/ros/jazzy/setup.bash \
    && source install/setup.bash \
    && ros2 launch kpi_rover launch_sim.launch.py"

Build and Launch Simulation Without Docker

Gazebo runs very slowly in Docker. Because of this, we need to run the project locally to make it work faster.

To avoid problems with missing packages, always add new packages to the package.xml file. This helps to install them easily using the rosdep tool.

Install ROS Dependencies

rosdep update
rosdep install --from-paths src --ignore-src -r -y

Source ROS2 (If Not Added to .bashrc)

source /opt/ros/jazzy/setup.bash

Build Project

colcon build

Launch Simulation

source install/setup.bash
ros2 launch kpi_rover launch_sim.launch.py

Build and Launch on RPI

Note: RPI, BBB and PC should be connected to the common network.

To set up automated startup on power-on(RPI):

Install udev rules for the lidar

cd ~/ros_ws/src/lidar_coin_d4a
sudo cp sc_mini.rules /etc/udev/rules.d

Build Docker

cd ~/ros_ws
docker build -t kpi-rover . -f docker/Dockerfile.rpi

Build Project

docker run --rm -it \
  --user $(id -u):$(id -g) \
  -v $(pwd):/workspace \
  -w /workspace \
  kpi-rover colcon build

Create system services

1. For DDS server to establish communication between nodes through the network

• Create /etc/systemd/system/fastdds.service system service file with the following contents. You need sudo permission to do this.

[Unit]
After=network-online.target
Description=DDS server for ROS2 node comunication through network

[Service]
User=vik
Group=vik
ExecStart=fastdds discovery -i 0
Restart=always
RestartSec=5s

[Install]
WantedBy=default.target

• Enable and start the service.

sudo systemctl enable fastdds.service
sudo systemctl start fastdds.service

2. For running ROS nodes

• Create /etc/systemd/system/kpi-rover.service system service file with the following contents. You need sudo permission to do this.

[Unit]
After=network-online.target
After=docker.service
Requires=docker.service
After=fastdds.service
Description=ROS2 autonomous robot services

[Service]
User=<user>
Group=<group>
ExecStart= bash -c "${HOME}/ros_ws/run.sh -b <bbb_ip>"
TimeoutStartSec=0
Restart=always
RestartSec=2s

[Install]
WantedBy=default.target

• Enable and start the service.

sudo systemctl enable kpi-rover.service
sudo systemctl start kpi-rover.service

To run visualization on your host machine:

export ROS_DISCOVERY_SERVER=<ip of the RPI>:11811

cd ~/ros_ws
source install/setup.bash
ros2 launch kpi_rover launch_visualization.launch.py 

Other Notes

This section holds temporary, unorganized notes.

Start Keyboard Control

ros2 run teleop_twist_keyboard teleop_twist_keyboard \
--ros-args \
-r /cmd_vel:=/diff_drive_base_controller/cmd_vel \
-p stamped:=True \
-p frame_id:=base_link

Launch CSPC Lidar Demo

ros2 launch cspc_lidar demo.launch.py

Start DDS Server

A DDS (Data Distribution Service) server facilitates real-time data exchange between distributed ROS2 nodes over the network. Using a DDS server ensures seamless communication between components, particularly when a PC has multiple network interfaces. It helps in optimizing discovery mechanisms and maintaining robust connectivity.

fastdds discovery -i 0

Launch in Docker when DDS is used

docker run --rm -it \
  --init \
  --user $(id -u):$(id -g) \
  -v $(pwd):/workspace \
  -w /workspace \
  -e ROS_DISCOVERY_SERVER=172.17.0.1:11811 \
  -e ROS_DOMAIN_ID=1 \
  -e IGN_DISCOVERY_URI=tcp://172.17.0.1:11811 \
  -e FASTRTPS_DEFAULT_PROFILES_FILE=/workspace/super_client_cfg_file.xml \
  -e GZ_PARTITION=1 \
  kpi-rover \
  bash -c "
    source /opt/ros/jazzy/setup.bash \
    && source install/setup.bash \
    && ros2 launch apricotka-robot-car launch_sim.launch.py"

TODO List

• Add entrypoint to surce source install/setup.bash
• Create Docker container for RPI
• How to use gamepad