Team_H

1. install educational version of CoppeliaSim_Edu_V4_1_0_Ubuntu20_04
2. install pyrep from https://github.com/stepjam/PyRep and follow installation quidelines

Presentation

https://docs.google.com/presentation/d/1AJWDzzrQ4JDr2ALND8X1Dvz7XjJeEq1TzlvenHl9Lks/edit?usp=sharing

Task Distribution

• Theresa Gräbner: Task 1 + Task 3 → Branch Theresa
• Peishi Liu: Task 2

References

• ARCL_23_24: team b
• https://github.com/stepjam/PyRep
• https://moveit.picknik.ai/humble/doc/tutorials/pick_and_place_with_moveit_task_constructor/pick_and_place_with_moveit_task_constructor.html

Tasks

• each Task has its own folder
• in /src you can find the executables
  • execute with "python Task[number].py"
• in files there are the results documented

Task 1

Uses CoppeliaSim_Edu_V4_1_0_Ubuntu20_04 and PyRep https://github.com/stepjam/PyRep

Task 1.1

• Goal: Test joint configurations for their validity
• Model: Given Franka Panda robot by chair

Implementation:

• Self-Collision check: check each link against all others except neighbouring links
• Write result in json format in “task1_1_self_collision.data”

Task 1.2

• Goal: test joint configurations and cartesian pose for validity + plan normalized/linear path if possible, otherwise plan an alternative path
• Model: Pyrep’s Panda Robot

Implementation:

• Scene without floor → not interested in collision with floor
• For every cartesian pose: Plan cartesian trajectory
• For every joint configuration: Plan joint trajectory

Cartesian Trajectroy Implementation:

• Load Cartesian Pose
• Extract quaternions and position coordinates via Pinocchio
• Test Cartesian Pose
  • Reachability: Sample valid joint configuration
  • Linear Trajectory: Pyrep’s get_linear_path
  • Nonlinear Trajectory: Pyrep’s get_nonlinear_path
• Write result in json format in “task1_2_cartesian_trajectory.data”

Joint Trajectory Implementation:

• Load joint configuration
• Reachability: Test if joint configurations are within the limits defined by Franka’s official website
• Validity: Configuration does not cause self-collision
• Normalized Trajectory:
  • Calculate trajectory
  • Validate each path point
• Alternative Path: Modified PyRep’s get_path
• Write result in json format in “task1_2_joint_trajectory.data”

Task 2: System Requirements and Setup Instructions

System Requirements

To complete this task, ensure your system meets the following requirements:

• Operating System: Ubuntu 22.04
• ROS 2 Distribution: Humble
• MoveIt 2 Version: Humble

Installation Instructions

Given that the required packages are quite numerous, it is recommended to follow the official MoveIt 2 installation guide for ROS 2 Humble. This guide will help you set up the necessary environment on your system.

Step-by-Step Guide

1. Follow the MoveIt 2 Humble Startup Instructions:

   • Visit the official MoveIt 2 Humble getting started guide for detailed instructions on installing and configuring MoveIt 2 on your system.

   • You can find the guide here.

   This guide covers everything from installing ROS 2 Humble to setting up MoveIt 2 and running the first demo.

2. build your workspace

   • cd ~/ws_moveit2/src and clone the Task2 pkg under branch Peishi
   • use this command to install all the dependenciesrosdep install --from-paths src --ignore-src -r -y
   • cd .. and build all the pkgs with colcon build
   • then source your workspace with source ~/ws_moveit2/install/setup.bash
   • start the simulation and visualize it with rviz, ros2 launch task2 Task2.launch.py

Task 2.1

• ros2 launch reach_and_grasp.launch.py

Task 2.2

• ros2 launch reach_and_grasp_cluttered.launch.py
• add the motion planning plugin into rviz and drag the button in the left bottom to see the Nullspace

Task 2.3

• not done

Task 3

Uses CoppeliaSim_Edu_V4_1_0_Ubuntu20_04 and PyRep https://github.com/stepjam/PyRep

Task 3.1

• Goal: Pick up cuboid and hold it with both robots
• Model: Pyrep’s Panda Robot
• Libraries: PyRep

Implementation:

• Move robots into starting position
• Right Robot → 0.55 | 0 | 0.01
• Left Robot → 0.7 | -0.1 | 0.3
• Right Robot grasps cuboid + moves to [0.7, -0.01, 0.3]
• Left Robot moves to [0.7, 0.01, 0.3] + grasps cuboid
• Save times and paths

Task 3.2

• Goal: Pick up Cuboid and move it in dual-mode along a circular path
• Model: Pyrep’s Panda Robot
• Libraries: PyRep

Execution:

• execute with "python Task3_2.py"
• programm asks if it should use old solution
  • write "y" for recreating the last solution
  • write "n" for recalculating solution

Why change configuration at every waypoint?

• corrects pose of robot
• adds stability
• moves robot further away from singular configurations, which makes it easier to calculate a functioning, stable subtrajectory
• if robot comes to configurations which are close to its limits, the cartesian trajectory can most likely not be continued that way but the robot needs to change its configuration in place to continue
• used ARCL_23_24: team b as inspirement

Initial Configuration

• saved in Initialisation_Conf.data

Implementation: circular path in dual mode

• Move robots to starting position
• Calculate waypoints every 30 degree
• For every waypoint:
  • Move robots to initialization configuration of waypoint
  • Calculated 4 intermediate waypoints
  • Pyrep’s get_path_from_cartesian_path

Future Improvements:

• implement special control algorithm, such that robots at 180 degrees don't have to interrupt dual mode.
• Robots should be able to move to new configuration without moving the end effector