/OMPLement

Simplifying OMPL for CoppeliaSim!

Primary LanguagePythonMIT LicenseMIT

OMPLement

Simplifying OMPL for CoppeliaSim (because sifting through the tutorials and forums is hard enough).

Overview

This is a simple script that you can add to your CoppeliaSim project to perform path planning. This is especially compatible with function calls through the remote API provided by CoppeliaSim.

How to Use?

  1. You will need to create a dummy object with a child script (feel free to call it what you want). In this child script, you should add everything found in the file OMPLement.py.

    • Alternatively, in this repository, I provide a blank scene with a dummy object called OMPLement. You can simply copy this object and paste it into your scene.

  2. From your code, you will need to call the path planning function from the script using sim.callScriptFunction, which requires:

    • the name of the function you wish to call from the child script (i.e., path_planning)
    • a script handle (from sim.getScript)
    • any number of arguments needed for the function.

  3. The path_planning function requires a single (Python) dictionary (or its Pythonic equivalent in whatever language you are using) as an argument. This dictionary must have the following entries:

    • robot - the name of the base of the robot in the CoppeliaSim scene
    • goal - the object handle of the target for the robot (perhaps for grasping)
    • algorithm - The name of the motion planning algorithm/solver to use (by default, RRTstar will be used); CoppeliaSim has several OMPL algorithms available.
    • arm_prefix (optional) - this is the name prefix given to joints and tip of the end-effector for which motion planning is being used.
      • By default, the joints use the format <robot_name>_joint<i>, where <robot> is the same string given as the robot argument above and <i> refers to the cardinality of the joints.
      • By default, the end-effector tip uses the format <robot_name>_tip.
      • For best use, you should make it that both the tip and joints of the arm share the same prefix!
    • num_attempts (default: 5) - the maximum number of iterations given for the solver to find a path plan. Each iteration by default tries to find a solution within 5 seconds. You should tune the arguments for the sim.compute function (line 162) as you need (e.g., increasing computation time).

  4. The function will return a list of configurations (referred to as path in this documentation).

    • If successful, path will be NxJ matrix (or 2D list), where N is the number of points in the path and J is the total number of joints. You would then need to iterate through this list while setting the configuration of the robot using sim.setJointPosition for each joint.
    • If no solution was found, path will be empty.

Your code in Python for instance may look like the following:

ompl_script = sim.getScript(sim.scripttype_childscript , sim.getObject('/OMPLement'))
path = sim.callScriptFunction(
	'path_planning',
	ompl_script,
	{
		"robot": <some_robot_name>,
		"goal": <target_obj>,
		"algorithm": <algorithm>,
		"num_attempts": <num>,
	},
)

Further Details on OMPL in CoppeliaSim

To deeply understand what's happening in the code, I would recommend going through the tutorials (yes, I know...) that are available online. There may be specific things you require for your own planning problem, so you should definitely consult the forums.

Questions? Comments? Feedback?

Feel free to reach out to me via email!