RCTA Manipulation

RCTA Manipulation is a set of catkin packages compatible with the ROS Indigo distribution on Ubuntu 14.04 that provide planning capabilities for manipulation on RCTA robotic platforms.

Building

Set up a catkin workspace as instructed in the ROS tutorials or determine an existing catkin workspace to host the rcta_manipulation packages.

Source dependencies are listed in rcta_manipulation/rcta.rosinstall and may be downloaded simultaneously via wstool. Alternatively, you can manually clone all the packages listed in that file and place them alongside the rcta_manipulation package in your catkin workspace.

System dependencies may be installed using rosdep. At the root of the workspace, executing the following command will prompt to install all required system dependencies:

rosdep install --from-paths rcta_manipulation -i -y

The recommended way to build the packages in this repository is to use the catkin from the python-catkin-tools package. A workspace containing these packages may be built using the catkin build command.

Running

There are multiple GNU screen configuration scripts in the roman_manipulation package for running the system under a few configurations. The system can be run under any configuration using rosrun. For example,

rosrun roman_manipulation screenrc_fake

will run the complete system with an offline (fake) simulation of the RoMan robot.

Executing a configuration script will start a screen session with tabs preloaded with the required commands. Launch the system by executing the command within each tab. Note that the default screen accelerator key C-a is modified to C-x. Ctrl + the arrow keys can be used to cycle through tabs.

Frequently used commands

Make sure that the time on robot's computer and time in the docker container/your computer are the same. The server is currently set in the /etc/hosts file to the IP of the robot.

sudo ntpdate -u tl1-1-am1 

To align the time zones in docker to the same as that on the robot, add this environment variable to the docker startup file.

--env="TZ=America/New_York" \

SSH into the robot computer

ssh rcta@tl1-1-am1

Run the export ROS_IP command on each of the docker tabs(set it to your computer's IP address). This is to inform the robot of the IP of the host computer.

Set the ROS_MASTER_URI if running the planner on the robot.

--env="ROS_MASTER_URI=http://tl1-1-am1:11311" \

If there is a need to manually move the robot's joints around, using the scripts from $RS_LIMB_ROOT/sbin.

Running visualization in RViz - Run this export LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64:$LD_LIBRARY_PATH to make sure RViz is running with the right libraries. The rviz file with the correct configs set is rcta-roman1.rviz. Run it with the namespace __ns:=roman1

Running batch tests on the robot -

While running the Realsense, launch the rs_rgbd.launch file with the it with the argument camera:=roman1/center_realsense

Moving the fingers

rostopic pub -1 /roman1/rcta_right_robotiq_controller/command std_msgs/Float64MultiArray 'layout:
  dim:
  - label: joint
    size: 4
    stride: 1
  data_offset: 0
data: [6.0, 6.0,6.0,137.0]'

Moving the dynamixel that supports the camera -

rcta@tl1-1-am1:~/rcta/install$ more share/rcta_dynamixel_publisher/config/roman1

To run PERCH

rostopic pub /requested_object std_msgs/String "data: 'test'"