demos

Demos

This directory contains examples and demonstration code in both C++ and Python that showcase the functionalities introduced in the different libraries of control-libraries.

Table of contents:

• Running a demo script
  • Task Space control loop example
  • Robot kinematics control loop example
• Other demos
  • ROS

Running a demo script

The easiest way to run them is to use the run-demo.sh file. Without arguments, this script creates a container and opens in interactive mode, allowing you to browse the different demo scripts and run the one of your choice:

./run-demo-script.sh
# Run a python script
root@xxxxxxxxx:/control_loop_examples$ python3 python_scripts/<script>.py
# Run a cpp script from the build folder
root@xxxxxxxxx:/control_loop_examples$ ./build/<script>

The available scripts are:

task_space_control_loop

• Showcased libraries: state_representation, dynamical_systems

This simple demonstration shows how to create a control loop with a PointAttractor dynamical system in task space (CartesianState). It moves a pose towards a random attractor in a 100Hz control loop. The script outputs the current pose and distance to the attractor at each timestep, and the final pose on reaching it:

...
frame CartesianPose expressed in world frame
position: (0.596701, 0.822976, -0.604581)
orientation: (0.246491, -0.314866, -0.896822, 0.189239) <=> theta: 2.64348, axis: (-0.32489, -0.925375, 0.195264)
distance to attractor: 0.001004
-----------
frame CartesianPose expressed in world frame
position: (0.596702, 0.822979, -0.604584)
orientation: (0.246488, -0.314866, -0.896823, 0.189239) <=> theta: 2.64348, axis: (-0.324891, -0.925375, 0.195264)
distance to attractor: 0.000994
-----------
##### TARGET #####
frame CartesianPose expressed in world frame
position: (0.59688, 0.823295, -0.604897)
orientation: (0.246242, -0.314924, -0.896867, 0.189256) <=> theta: 2.64399, axis: (-0.324929, -0.92536, 0.195269)
##### CURRENT POSE #####
frame CartesianPose expressed in world frame
position: (0.596702, 0.822979, -0.604584)
orientation: (0.246488, -0.314866, -0.896823, 0.189239) <=> theta: 2.64348, axis: (-0.324891, -0.925375, 0.195264)

robot_kinematics_control_loop

• Showcased libraries: state_representation, dynamical_systems, robot_model

This demonstration reuses the previous dynamical system control loop but adds the robot component. The desired command, i.e. desired twist of the robot end-effector (eef) is sent to a dummy robot. At each timestep, the new joint state and eef state of the robot is computed using the robot kinematics from the robot model corresponding to the provided URDF. The robot is assumed to perfectly follow the computed desired state matching the eef desired twist. Similarly to the previous demonstration, the script outputs the current joint positions, eef pose and distance to the attractor at each timestep, and the final joint positions and pose on reaching it:

franka JointPositions
names: [panda_joint1, panda_joint2, panda_joint3, panda_joint4, panda_joint5, panda_joint6, panda_joint7, ]
positions: [-1.39453, 0.34338, 1.78372, -1.21918, -0.365912, 1.1687, -2.88819, ]
panda_link8 CartesianPose expressed in panda_link0 frame
position: (0.500692, 4.30377e-05, 0.750494)
orientation: (-9.36521e-05, 8.70069e-05, 1, -0.000152405) <=> theta: 3.14141, axis: (-8.70069e-05, -1, 0.000152405)
distance to attractor: 0.001249
-----------
franka JointPositions
names: [panda_joint1, panda_joint2, panda_joint3, panda_joint4, panda_joint5, panda_joint6, panda_joint7, ]
positions: [-1.39627, 0.352394, 1.79511, -1.2227, -0.374677, 1.16757, -2.88157, ]
panda_link8 CartesianPose expressed in panda_link0 frame
position: (0.500192, 1.07333e-05, 0.750183)
orientation: (-1.9014e-05, 4.77102e-05, 1, -7.96025e-05) <=> theta: 3.14155, axis: (-4.77102e-05, -1, 7.96025e-05)
distance to attractor: 0.000455
-----------
##### TARGET #####
panda_link8 CartesianPose expressed in panda_link0 frame
position: (0.5, 0, 0.75)
orientation: (6.12323e-17, 0, 1, 0) <=> theta: 3.14159, axis: (0, 1, 0)
##### CURRENT STATES #####
franka JointPositions
names: [panda_joint1, panda_joint2, panda_joint3, panda_joint4, panda_joint5, panda_joint6, panda_joint7, ]
positions: [-1.39215, 0.362964, 1.80315, -1.22376, -0.385704, 1.16489, -2.87388, ]
panda_link8 CartesianPose expressed in panda_link0 frame
position: (0.500192, 1.07333e-05, 0.750183)
orientation: (-1.9014e-05, 4.77102e-05, 1, -7.96025e-05) <=> theta: 3.14155, axis: (-4.77102e-05, -1, 7.96025e-05)

Other demos

This section contains links to other demos and examples in the community. If you have an example to share, post it on Control Libraries Discussions and we will link it here.

ROS

The following examples by domire8 show how control libraries can be integrated with the Robot Operating System (ROS)

• https://github.com/domire8/control-libraries-ros-demos