Planning with the UR5 and the IGM Package
When the system has been already set it up and running the next step would be to plan a trajectory using RViz plug-in, the move_group_interface (C++) or the moveit_commander (python).
For this propose, it exists a package which can be used in order to plan trajectories using the move_group_interface. This package is still available under the
IGM repository. Clone this repository in your src folder in your catkin_ws directory and build your catkin space again.
The following are methods used by Moveit in order to specify a desired pose in the space:
- Joint space
- Constraint method
- Pose method
- Cartesian path (using Waypoints)
In the first case the positions, velocities and accelerations of the different joints can be specify following the order:
['shoulder_pan_joint', 'shoulder_lift_joint', 'elbow_joint', 'wrist_1_joint', 'wrist_2_joint', 'wrist_3_joint']
In order to get further information please visit the documentation pages of the project in \cite{rosproject}.
Object Avoidance
One of the most important things when path planning is to avoid object which could appear in the workspace of the Robot we are working on. In this case, it is necessary to account with an algorithm which can carry out object detection operations and object avoidance, furthermore as a part of the COAR project the following architecture has been built in order to simplify and do more effective the object recognition processes and its integration in the Path planning problem.
As it is shown, the main issue has been divided into two parts, fixed objects and random objects. in the first case, a node from where different kind of fixed objects in the workspace could be declared is created, beside this, another node similar to the last one is created. The difference between these nodes is that The last one is going to be mapping with a hight frequency such that if a new object has been attached or added to the workspace the system would be able to detect it and integrate this object into the object avoidance procedure. Because of this, the computer performance would not be compromised, however the execution time of the different tasks.
Universal Robot 5 Manual by Jorge De La Cruz is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.