Skip to content

Kinovarobotics/ros2_kortex

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ROS 2 Kortex

Kinova® Kortex™ is the common software platform behind all of the products in the Gen3 family (Gen3 and Gen3 lite). It unifies the inner workings of the various robots and their related external tools, like the API.
https://www.kinovarobotics.com/product/gen3-robots

Kinova Gen3 7DoF manipulator with Intel RealSense 3D Vision Module and Robotiq 2F-85 2 Finger 85mm Adaptive Gripper

ROS2 Kortex is the official ROS2 package to interact with Kortex and its related products. It is built upon the Kortex API, documentation for which can be found in the GitHub Kortex repository.

Build status

ROS 2 Distro Humble Iron Rolling
Branch main main main
Build Status Humble Binary Build Iron Binary Build Rolling Binary Build
Release Status coming soon coming soon coming soon

Note: There are several CI jobs checking against future upstream changes see detailed build status for a full list of CI jobs and for more information.

Getting started

  1. Install ROS 2.

    If you're a developer, we recommend using Rolling to get the latest features and fixes.

    Rolling Release: Install ROS2 Rolling
    Latest Release: Install ROS2 Iron
    Stable LTS Release: Install ROS2 Humble

    After installing a version of ROS, source the setup.bash, which will set the $ROS_DISTRO environment variable.

  2. Install this package from binary

    sudo apt install ros-$ROS_DISTRO-kortex-bringup
    
  3. Optional: install MoveIt Configuration and Cyclone DDS

    If you have a 7dof arm:

    sudo apt install ros-$ROS_DISTRO-kinova-gen3-7dof-robotiq-2f-85-moveit-config
    

    If you have a 6dof arm:

    sudo apt install ros-$ROS_DISTRO-kinova-gen3-6dof-robotiq-2f-85-moveit-config
    

    If you plan to use MoveIt, it is recommended to install and use Cyclone DDS.

    sudo apt install ros-$ROS_DISTRO-rmw-cyclonedds-cpp
    export RMW_IMPLEMENTATION=rmw_cyclonedds_cpp
    
  4. Go to Usage section

Contributing to this repository or building from source

Note: It is recommended to use a released binary version of this package and apt install it. If you want the latest version of this repository for testing latest fixes check out testing with pre-released binaries: https://docs.ros.org/en/rolling/Installation/Testing.html

If the bug fix you need isn't in a released version or If you want to build this repository from source or contribute back to the repository read on.

  1. Make sure that colcon, its extensions, and vcs are installed:

    sudo apt install python3-colcon-common-extensions python3-vcstool
    
  2. Create a new ROS2 workspace:

    export COLCON_WS=~/workspace/ros2_kortex_ws
    mkdir -p $COLCON_WS/src
    
  3. Pull relevant packages:

    cd $COLCON_WS
    git clone https://github.com/Kinovarobotics/ros2_kortex.git src/ros2_kortex
    vcs import src --skip-existing --input src/ros2_kortex/ros2_kortex.$ROS_DISTRO.repos
    vcs import src --skip-existing --input src/ros2_kortex/ros2_kortex-not-released.$ROS_DISTRO.repos
    

    If you plan on simulating the robot with ignition or gazebo, make sure to pull the additional simulation packages. If you're on ROS2 Humble, run

    vcs import src --skip-existing --input src/ros2_kortex/simulation.humble.repos
    

    otherwise

    vcs import --skip-existing --input src/ros2_kortex/simulation.repos
    

    If you plan on using MoveIt, you must make sure that you have it already installed either from binaries or by building it from source.

    If you plan on simulating the Gen3 7Dof robot mounted on the Husky mobile robot from clearpath, make sure to pull the additional related packages. On ROS2 Humble, run

    vcs import src --skip-existing --input src/ros2_kortex/clearpath.repos
    
  4. Install dependencies, compile, and source the workspace:

    rosdep install --ignore-src --from-paths src -y -r
    colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release
    

    By default, colcon will use as much resources as possible to build the ROS2 workspace. This can temporarily freeze or even crash your machine. You can limit the number of threads used to avoid this issue, we found a good tradeoff between build time and resource utilisation by setting it to 3 :

    colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release --parallel-workers 3
    
  5. Source the previously built workspace using the following command:

    echo 'source ~/workspace/ros2_kortex_ws/install/setup.bash' >> ~/.bashrc
    

Simulation Issues

Please note, at this time there are two known issues you with simulation

  1. Gazebo + Mimic Joints for the Robotiq Gripper
  2. Protobuf version mismatch

Gazebo and Mimic Joints

A pull request has been made to gz_ros2_control which is how this repository was tested in simulation. The pull request won't be merged as the fix should be done upstream in gz-sim. Once a fix is available ros2_robotiq_gripper will be re-released and an update should fix any workarounds.

In the meantime if you need simulation checkout the upstream pull request link:

Protobuf

Due to mismatched protobuf version that ships system and used by Gazebo simulator compiling twice may be required. You will only run into this if you have certain other gazebo related code in your workspace while compiling this repository. If errors are encounter you must clean your workspace and run colcon build in two steps:

  1. build everything except kortex related packages
  2. build the packages that where skipped
sudo apt install python3-colcon-clean # if you don't have colcon-clean installed already
colcon clean workspace -y
colcon build --packages-skip-regex '.*kortex.*' '.*gen3.*'
colcon build --packages-select-regex '.*kortex.*' '.*gen3.*'

Usage

To launch and view any of the robot's URDF run:

ros2 launch kortex_description view_robot.launch.py

The accepted arguments are:

  • robot_type : Your robot model. Possible values are either gen3 or gen3_lite, the default is gen3.

  • gripper : Gripper to use. Possible values for the Gen3 are either robotiq_2f_85 or robotiq_2f_140. For the Gen3 Lite, the only option is gen3_lite_2f. Default value is an empty string, which will display the arm without a gripper.

  • dof : Degrees of freedom of the arm. Possible values for the Gen3 are either 6 or 7. For the Gen3 Lite, the only option is 6. Default value is 7.

Gen 3 Robots

The gen3.launch.py launch file is designed to be used for Gen3 arms. The typical use case to bringup and visualize the 7 DoF Kinova Gen3 robot arm (default) with mock hardware on Rviz:

ros2 launch kortex_bringup gen3.launch.py \
  robot_ip:=yyy.yyy.yyy.yyy \
  use_fake_hardware:=true

Alternatively, for a physical robot:

ros2 launch kortex_bringup gen3.launch.py \
  robot_ip:=192.168.1.10 \

You can specify the following arguments if you wish to change your arm configuration:

  • robot_type: Your robot model. Default value (and only one) is gen3.

  • gripper : Gripper to use. Possible values for the Gen3 are either robotiq_2f_85, robotiq_2f_140 or "". Default is robotiq_2f_85. An empty string will not initialise any gripper.

  • gripper_joint_name : Name of the controlled joint of the gripper attached to the arm. Default value is robotiq_85_left_knuckle_joint.

  • use_internal_bus_gripper_comm : Use internal bus for gripper communication. Default value is true.

  • gripper_max_velocity : Max velocity for gripper commands. Default value is 100.0.

  • gripper_max_force : Max force for gripper commands. Default value is 100.0.

  • dof : Degrees of freedom of the arm. Possible values are either 6 or 7.Default value is 7.

  • robot_ip : IP address by which the robot can be reached. No default is specified, this is a required argument. All arms are shipped with address 192.168.1.10, but if you have reassigned your physical arm's robot IP address, then you will need to assign that ip address.

  • use_fake_hardware : Start robot with fake hardware mirroring command to its states. Default value is false.

  • fake_sensor_commands : Enable fake command interfaces for sensors used for simple simulations. Used only if 'use_fake_hardware' parameter is true. Default value is false.

  • robot_controller : Robot controller to start. Possible values are twist_controller and joint_trajectory_controller.Default value is joint_trajectory_controller.

  • controllers_file : Ros 2 control configuration file to use. Default value is ros2_controllers.yaml

  • launch_rviz : Start an Rviz window to visualize the robot. Default value is true.

Robotiq gripper

The Robotiq 2f 85 (or 2f 140) Gripper will be available on the Action topic:

/robotiq_gripper_controller/gripper_cmd

You can test the gripper by calling the Action server with the following command and setting the desired position of the gripper (0.0=open, 0.8=close)

ros2 action send_goal /robotiq_gripper_controller/gripper_cmd control_msgs/action/GripperCommand "{command:{position: 0.0, max_effort: 100.0}}"

Vision Module

In order to access the Kinova Vision module's depth and color streams for the camera-equipped Gen3 arm models, please refer to the following github repository for detailed instructions: ros2_kortex_vision

While following the instructions, please take note of the following points:

  1. There is no need to install the rgbd_launch ROS package
  2. Establishing a connection between the computer and the camera may require several attempts, so please be patient. Sometimes you may need to restart both the robot and the connected computer to successfully establish the connection.
  3. Before setting the depth_registration argument to true in the kinova_vision.launch.py file, make sure to install the image_proc ROS package on your system using the following command:
sudo apt install ros-$ROS_DISTRO-depth-image-proc
  1. After starting the kinova_vision.launch.py file, open RViz and add the desired camera topics to visualize the captured scene.

Gen 3 Lite Robot

The gen3_lite.launch.py launch file is designed to be used for Gen3 Lite arms. The typical use case to bringup the robot arm with mock hardware:

ros2 launch kortex_bringup gen3_lite.launch.py \
  robot_ip:=yyy.yyy.yyy.yyy \
  use_fake_hardware:=true

Alternatively, if you wish to use the physical robot:

ros2 launch kortex_bringup gen3_lite.launch.py \
  robot_ip:=192.168.1.10 \

You can specify the following arguments if you wish to change your arm configuration:

  • robot_type: Your robot model. Default value (and only one) is gen3_lite.

  • gripper : Gripper to use. Default value (and only one) is gen3_lite_2f.

  • gripper_joint_name : Name of the controlled joint of the gripper attached to the arm. Default value (and only one) is right_finger_bottom_joint.

  • use_internal_bus_gripper_comm : Use internal bus for gripper communication. Default value is true.

  • gripper_max_velocity : Max velocity for gripper commands. Default value is 100.0.

  • gripper_max_force : Max force for gripper commands. Default value is 100.0.

  • robot_ip : IP address by which the robot can be reached. No default is specified, this is a required argument. All arms are shipped with address 192.168.1.10, but if you have reassigned your physical arm's robot IP address, then you will need to assign that ip address. If you're using an USB to Ethernet interface to connect your robot to your machine instead of USB via RNDIS, the ip address will be 192.168.2.10.

  • use_fake_hardware : Start robot with fake hardware mirroring command to its states. Default value is false.

  • fake_sensor_commands : Enable fake command interfaces for sensors used for simple simulations. Used only if 'use_fake_hardware' parameter is true. Default value is false.

  • robot_controller : Robot controller to start. Possible values are twist_controller and joint_trajectory_controller.Default value is joint_trajectory_controller.

  • controllers_file : Ros 2 control configuration file to use. Default value is ros2_controllers.yaml

  • description_file : URDF/XACRO description file with the robot. Default value is gen3_lite_gen3_lite_2f.xacro.

  • launch_rviz : Start an Rviz window to visualize the robot. Default value is true.

Simulation

The kortex_sim_control.launch.py launch file is designed to simulate all of our arm models, you just need to specify your configuration through the arguments. By default, the Gen3 7 dof configuration is used :

ros2 launch kortex_bringup kortex_sim_control.launch.py \
  use_sim_time:=true \
  launch_rviz:=false
  • sim_ignition : Use Ignition for simulation. Default value is true.
  • sim_gazebo : Use Gazebo Classic for simulation. Default value is false.
  • robot_type : Your robot model. Possible values are either gen3 or gen3_lite.Default is gen3.
  • robot_name : Name you would like your robot to have. Default value is gen3.
  • dof : Degrees of freedom of the arm. Possible values are either 6 or 7.Default value is 7.
  • vision : Use arm mounted realsens. Possible values are either true or false. Default value is false. This option does not generate simulated images, it only loads up the robot's URDF that includes the vision link.
  • robot_controller : Robot joint controller to start. Default value is joint_trajectory_controller.
  • robot_pos_controller : Robot position controller to start. Default value is twist_controller.
  • robot_hand_controller : Robot gripper controller to start. Default value is robotiq_gripper_controller.
  • controllers_file : Ros 2 control configuration file to use. Default value is ros2_controllers.yaml
  • description_package : Description package with robot URDF/XACRO files. Default value is kortex_description.
  • description_file : URDF/XACRO description file with the robot. Default value is kinova.urdf.xacro.
  • prefix : Prefix of the joint names, useful for multi-robot setup. If changed, then also joint names in the controllers' configuration have to be updated. Default value is "" (none).
  • use_sim_time : Use simulated clock. Default value is true.
  • gripper : Gripper to use. Possible values for the Gen3 are either robotiq_2f_85, robotiq_2f_140 or "". Default is robotiq_2f_85. An empty string will not initialise any gripper.

MoveIt2

To generate motion plans and execute them with a simulated 7 DoF Kinova Gen3 arm with mock hardware:

ros2 launch kinova_gen3_7dof_robotiq_2f_85_moveit_config robot.launch.py \
  robot_ip:=yyy.yyy.yyy.yyy \
  use_fake_hardware:=true

and to bring up the Kinova Gen3 6 DoF with MoveIt:

ros2 launch kinova_gen3_6dof_robotiq_2f_85_moveit_config robot.launch.py \
  robot_ip:=yyy.yyy.yyy.yyy \
  use_fake_hardware:=true

To generate motion plans and execute them with an ignition simulated 7 DoF Kinova Gen3 arm (previously launched with the command at the simulation section):

ros2 launch kinova_gen3_7dof_robotiq_2f_85_moveit_config sim.launch.py \
  use_sim_time:=true

To work with a physical robot and generate/execute paths with MoveIt run the following:

For Gen3:

ros2 launch kinova_gen3_7dof_robotiq_2f_85_moveit_config robot.launch.py \
  robot_ip:=192.168.1.10

For Gen3-Lite:

ros2 launch kinova_gen3_lite_moveit_config robot.launch.py \
  robot_ip:=192.168.1.10

Commanding the arm (physically and in simulation)

You can command the arm by publishing Joint Trajectory messages directly to the joint trajectory controller:

ros2 topic pub /joint_trajectory_controller/joint_trajectory trajectory_msgs/JointTrajectory "{
  joint_names: [joint_1, joint_2, joint_3, joint_4, joint_5, joint_6, joint_7],
  points: [
    { positions: [0, 0, 0, 0, 0, 0, 0], time_from_start: { sec: 10 } },
  ]
}" -1

Depending on your robot type and its DoF, you will need to adapt the joint_names and positions properties accordingly. For the Gen3 Lite arm, the integrated gripper is considered as a joint, so to command it, it must be included in the joint_names array. (0.0=open, 1.0=close):

ros2 topic pub /joint_trajectory_controller/joint_trajectory trajectory_msgs/JointTrajectory "{
  joint_names: [joint_1, joint_2, joint_3, joint_4, joint_5, joint_6, right_finger_bottom_joint],
  points: [
    { positions: [0, 0, 0, 0, 0, 0, 1], time_from_start: { sec: 10 } },
  ]
}" -1

You can also command the arm using Twist messages. Before doing so, you must active the twist_controller and deactivate the joint_trajectory_controller:

ros2 service call /controller_manager/switch_controller controller_manager_msgs/srv/SwitchController "{
  activate_controllers: [twist_controller],
  deactivate_controllers: [joint_trajectory_controller],
  strictness: 1,
  activate_asap: true,
}"

Note: the required interface for the twist_controller does not currently exist in the gazebo or mock hardware simulation setups. So the twist_controller is currently only functional on Kinova hardware.

Once the twist_controller is activated, You can publish Twist messages on the /twist_controller/commands topic to command the arm.

For example, you can jog the arm using Teleop Twist Keyboard with the following command:

WARNING: you are responsible for collision checking, including self collisions when in this mode.

ros2 run teleop_twist_keyboard teleop_twist_keyboard --ros-args --remap /cmd_vel:=/twist_controller/commands

If you wish to use the joint_trajectory_controller again to command the arm using JointTrajectory messages, run the following:

ros2 service call /controller_manager/switch_controller controller_manager_msgs/srv/SwitchController "{
  activate_controllers: [joint_trajectory_controller],
  deactivate_controllers: [twist_controller],
  strictness: 1,
  activate_asap: true,
}"

Contents

The following is a description of the packages included in this repository.

kortex_description

This package contains the URDF (Unified Robot Description Format), STL and configuration files for the Kortex-compatible robots. For more details, please consult the README from the package subdirectory.

kortex_driver

This package implements a ROS node that allows communication between a node and a Kinova Gen3 or Gen3 lite robot. For more details, please consult the README from the package subdirectory.

kortex_moveit_config

This metapackage contains the auto-generated MoveIt! files to use the Kinova Gen3 and Gen3 lite arms with the MoveIt! motion planning framework. For more details, please consult the README from the package subdirectory.