Since this repo contains a submodule, barrett_model, you must clone using the following command (as described in the git documentation)

git clone --recurse-submodules https://github.com/saszaz/barrett_wam_gazebo_sim

To change to the project-relevant branch for this repo:

cd barrett_model
git checkout --track origin/my_devel

The barrett_wam.launch file will start the Gazebo simulator, load the urdf model, and start the controller interfaces. The ctrl_mode argument can be set to position (default), velocity, or effort. The robot state, including joint values and tf transforms are published on corresponding rostopics.

roslaunch barrett_wam_gazebo barrett_wam.launch ctrl_mode:=position

This will start the move_group planning interface, and allow path planning using the MoveIt! library.

roslaunch barrett_wam_gazebo barrett_wam_moveit_planning_execution.launch debug:=true sim:=true

This includes a manual user-interface for setting start/goal positions of the robot end-effector, and generating plans from selected OMPL path planners.

roslaunch barrett_wam_moveit_config moveit_rviz.launch config:=true

Within the barrett_wam_moveit_control package, you can generate an elliptical trajectory using the matlab file scripts/path_generator.m and making sure to set the appropriate parameters. This will create a cart_path_ellipse_n_***.csv under the trajectories directory.

Currently, the .csv file to use for trajectory following can be set in src/move_group_interface.cpp. TO-DO: read this selection (along with other parameters) from a config file to avoid re-compilation.

You can then run the waypoint-following script through the launch file:

roslaunch barrett_wam_moveit_control moveit_control.launch

The arm will attempt to move to a pre-set start configuration, and execute the trajectory.