Kinova Jaco 2 Arm Docker

Dockerfile for running Kinova Jaco 2 without any ROS installed on the machine. Tested on Ubuntu 18.04.

Instructions

• Install Docker from here, and follow the instructions here to run docker without sudo access.

• Pull the ros:melodic container from Docker

docker pull ros:melodic

• Clone this repo into ~/workspace and build the docker image:

mkdir ~/workspace && cd ~/workspace
git clone https://github.com/melfm/jaco_docker.git
cd jaco_docker
docker build --tag jaco_control .

• To access the arm via USB copy the udev srule file 10-kinova-arm.rules from /udev to /etc/udev/rules.d and then connect the robot to your machine:

sudo cp udev/10-kinova-arm.rules /etc/udev/rules.d/

• Run the Docker container with access to the USB devices for establishing connection to the robot:

docker run -it --name jaco_robot_net --privileged -v /dev/bus/usb:/dev/bus/usb  -p 9030:9030 jaco_control

• You can now run the robot from inside the Docker regularly as if you had ROS installed on your machine. In order to connect to the running container use the following command:

docker exec -it jaco_robot_net /bin/bash

• Hint, if docker exec doesn't work, you may need to run docker start on the container id (found with docker ps)

• To build the catkin workspace

catkin_make

• To remove existing container and launch a new one

docker rm "name_of_container"

RealSense Camera

Instructions from Docker is fine but on my laptop I have also tried these steps in the past to get the camera working:

https://github.com/IntelRealSense/realsense-ros#step-2-install-intel-realsense-ros-from-sources

Just to be sure:

echo "source ~/catkin_ws/devel/setup.bash" >> ~/.bashrc
source ~/.bashrc

To launch the camera:

roslaunch realsense2_camera rs_camera.launch

Any problems with not detecting the camera, is usually related to blocked port on the computer. Usually resetting the computer and plugging the camera before seems to work - no real solution found for this.

Now Launch rviz (you would need to run the container via the script run_docker_image.bash instead:

rosrun rviz rviz

Set the Reference Frame to camera_color_frame and add a Camera object and set the topic to /camera/color/image_raw - You should see the image frames.