rxt_skills_chasi

Implementation of ROBxTASK skills for ARTI Chasi robot as ROS Action Server

Start and drive with the robot

• Long press the left Button on the ARTI Robot to switch it on
• Connect the computer to the ARTI Chasi Wi-Fi
• Open a Terminal and connect to the ARTI Robot (Rpi4) via ssh [email protected] - password: ubuntu
• Start the ros nodes on the Robot with roslaunch arti_chasi_mark3 arti_chasi_mark3_upstart_with_teleop.launch
• Open a second Terminal and start the OS1 lidar and Rviz with the launch file in this repository roslaunch arti_navigation arti_with_os1.launch
• Switch on the Xbox 360 Controller
• Drive!, but be careful, the robot reacts very fast and has a high acceleration

Joystick Control on XBOX Controller:

• LB → Deadman switch (has to be pressed all the time for control)
• Left Joystick → Drive
• Right Joystick → Steer

Initial Setup - Creating a ROS Workspace

• Install Ubuntu 16.04
• Install ROS Kinetic (install ros-kinetic-desktop-full)
• Install Gazebo 7.15+
• Create catkin workspace
• Follow Instructions on https://github.com/ARTI-Robots/chasi
• Clone https://github.com/ARTI-Robots/ackermann-drive-teleop into catkin_workspace/src

Next we need to get some packages:

• sudo apt-get install ros-kinetic-controller-manager
• sudo apt install ros-kinetic-gazebo-ros-control
• sudo apt install ros-kinetic-effort-controllers
• sudo apt install ros-kinetic-joint-state-controller
• (recommended) install the complete ros_control package including joint controllers (https://wiki.ros.org/ros_control)
• sudo apt install ros-kinetic-ros-control ros-kinetic-ros-controllers
• (recommended) install moveit for motion planning (https://moveit.ros.org/install/)
• sudo apt install ros-kinetic-moveit
• catkin make inside catkin_workspace

Simulation with a virtual Robot:

• Connect Gamepad to PC via Cable
• Start Rviz for visualization and Gazebo for simulation → roslaunch arti_chasi_gazebo gazebo_with_ouster_16.launch
• Start the gamepad control node → roslaunch ackermann_drive_teleop ackermann_drive_joyop.launch

Network Configuration

For Wifi-Configuration see also: ARTI Herstellerinformation / Handbuch

• SSID: "ARTI Chasi", Password: ARTIDefaultPW1!
• Configuration: https://192.168.5.1/ (User & Password "ubnt")
• EdgeMAX Router IP (DHCP Server): 192.168.5.1
• Raspberry PI IP: 192.168.5.3
• Ouster LIDAR IP: 192.168.5.4
• Notebook IP: 192.168.5.5

Running the LIDAR (on the notebook):

For install:

• ROS Workspace should already be created on the notebook
• Install the ouster-ros package into the workspace
• cd ~/arti_ws/src
• git clone https://github.com/ouster-lidar/ouster_example.git
• export CMAKE_PREFIX_PATH=~/arti_ws/src/ouster_example
• cd ~/arti_ws
• catkin_make -DCMAKE_BUILD_TYPE=Release

To Run:

• Start the Ouster LIDAR via ROS → roslaunch ~/arti_ws/src/ouster_example/ouster_ros/os1.launch os1_hostname:=192.168.5.4 os1_udp_dest:=192.168.5.5 lidar_mode:=1024x20 viz:=false
• For different configuration parameters see: Ouster GitHub
• Visualize the Data with rviz → rviz -d ~/arti_ws/src/ouster_example/ouster_ros/viz.rviz

Synchronize Time between Machines and LIDAR

This is necessary in order for the SLAM Algorithms to work properly.

For the synchronisation between the ARTI Rpi and the Notebook:

• Install chrony on both machines
• apt-get install chrony
• Configure the NTP Server (Notebook)
• sudo nano /etc/chrony/chrony.conf
• Add line: "local stratum 8"
• Add line: "allow 192.168.5"
• Configuire the NTP Client (ARTI Rpi)
• sudo nano /etc/chrony/chrony.conf
• Add the lines: server 192.168.5.5 minpoll 0 maxpoll 5 maxdelay .05
• Reboot both machines
• Verify the synchronization on the client with: chronyc tracking (should show the IP of the notebook and show its synchronized).

Map Making and Localization (SLAM) with Google Cartographer:

Important! Cartographer is not buildable now. First, the github repo states "build failed" for a long time. Second ROS Kinetic is not supported anymore by Cartographer, since its EOL. I tried to create a own rosinstall file (see arti_ws/src/arti_navigation/scripts/install cartographer.sh) and replaced the "version master" with two older commits which did not fail building at the CI pipeline of cartographer, but that also doesnt work. Only solution until now: Copy the cartographer_ws from the old notebook, delete the build_isolated and install_isolated folders and then run catkin_make_build --install --ninja only (see arti_ws/src/arti_navigation/scripts/install cartographer.sh). A backup of the cartographer_ws is located at my backup drive. (and on the new and old ARTI Notebook)

see GitHub: https://github.com/nerovalerius/arti_navigation

Prerequisites

• Install google cartographer with rosrun arti_navigation install_cartographer.sh - this This installs as described "here"
• Use rosrun arti_navigation configure_terminals.sh to add the necessary lines to bashrc which source the cartographer_ws and the arti_ws
• Start the LIDAR with an visualization
• Connect the computer to the ARTI Chasi Wi-Fi
• Open a Terminal and connect to the ARTI Robot (Rpi4) via ssh [email protected] - password: ubuntu
• Start the ros nodes on the Robot with roslaunch arti_chasi_mark3 arti_chasi_mark3_upstart_with_teleop.launch
• Open a second Terminal and start the OS1 lidar and Rviz with the launch file in this repository roslaunch arti_navigation arti_with_os1.launch
• Switch on the Controller
• Start the map making process with google cartographer
• After the ros nodes on the robot and the os1_lidar with rviz is launched, start the cartograhper ROS node with roslaunch arti_navigation arti_cartographer.launch
• Inside Rviz, the LIDAR data can be visualized with Add → By Topic → PointCloud2. Also the color scheme of the LIDAR data can be manipulated for easier interpretation.