Update tutorials

Author: sjahr

CMakeLists.txt

@@ -69,6 +69,7 @@ add_subdirectory(doc/how_to_guides/parallel_planning)
 add_subdirectory(doc/how_to_guides/chomp_planner)
 add_subdirectory(doc/how_to_guides/persistent_scenes_and_states)
 add_subdirectory(doc/how_to_guides/pilz_industrial_motion_planner)
+add_subdirectory(doc/how_to_guides/pipeline_testbench)
 add_subdirectory(doc/how_to_guides/using_ompl_constrained_planning)
 add_subdirectory(doc/tutorials/pick_and_place_with_moveit_task_constructor)
 add_subdirectory(doc/tutorials/quickstart_in_rviz)

doc/concepts/motion_planning.rst

@@ -47,36 +47,30 @@ Pre-processing is useful in several situations, e.g. when a start state for the
 Post-processing is needed for several other operations, e.g. to convert paths generated for a robot into time-parameterized trajectories.
 MoveIt provides a set of default motion planning adapters that each perform a very specific function.
 
-FixStartStateBounds
+CheckStartStateBounds
 ^^^^^^^^^^^^^^^^^^^
 
 The fix start state bounds adapter fixes the start state to be within the joint limits specified in the URDF.
 The need for this adapter arises in situations where the joint limits for the physical robot are not properly configured.
 The robot may then end up in a configuration where one or more of its joints is slightly outside its joint limits.
 In this case, the motion planner is unable to plan since it will think that the starting state is outside joint limits.
-The "FixStartStateBounds" planning request adapter will "fix" the start state by moving it to the joint limit.
+The "CheckStartStateBounds" planning request adapter will "fix" the start state by moving it to the joint limit.
 However, this is obviously not the right solution every time - e.g. where the joint is really outside its joint limits by a large amount.
 A parameter for the adapter specifies how much the joint can be outside its limits for it to be "fixable".
 
-FixWorkspaceBounds
+ValidateWorkspaceBounds
 ^^^^^^^^^^^^^^^^^^
 
 The fix workspace bounds adapter will specify a default workspace for planning: a cube of size 10 m x 10 m x 10 m.
 This workspace will only be specified if the planning request to the planner does not have these fields filled in.
 
-FixStartStateCollision
+CheckStartStateCollision
 ^^^^^^^^^^^^^^^^^^^^^^
 
 The fix start state collision adapter will attempt to sample a new collision-free configuration near a specified configuration (in collision) by perturbing the joint values by a small amount.
 The amount that it will perturb the values by is specified by the **jiggle_fraction** parameter that controls the perturbation as a percentage of the total range of motion for the joint.
 The other parameter for this adapter specifies how many random perturbations the adapter will sample before giving up.
 
-FixStartStatePathConstraints
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-This adapter is applied when the start state for a motion plan does not obey the specified path constraints.
-It will attempt to plan a path between the current configuration of the robot to a new location where the path constraint is obeyed.
-The new location will serve as the start state for planning.
 
 AddTimeParameterization
 ^^^^^^^^^^^^^^^^^^^^^^^

doc/examples/motion_planning_pipeline/src/motion_planning_pipeline_tutorial.cpp

@@ -141,6 +141,8 @@ int main(int argc, char** argv)
   req.pipeline_id = "ompl";
   req.planner_id = "RRTConnectkConfigDefault";
   req.allowed_planning_time = 1.0;
+  req.max_velocity_scaling_factor = 1.0;
+  req.max_acceleration_scaling_factor = 1.0;
   planning_interface::MotionPlanResponse res;
   geometry_msgs::msg::PoseStamped pose;
   pose.header.frame_id = "panda_link0";

doc/examples/moveit_cpp/config/moveit_cpp.yaml

@@ -9,10 +9,11 @@ planning_scene_monitor_options:
 
 planning_pipelines:
   #namespace: "moveit_cpp"  # optional, default is ~
-  pipeline_names: ["ompl"]
+  pipeline_names: ["stomp"]
 
 plan_request_params:
   planning_attempts: 1
-  planning_pipeline: ompl
+  planning_pipeline: stomp
+  planner_id: "stomp"
   max_velocity_scaling_factor: 1.0
   max_acceleration_scaling_factor: 1.0

doc/examples/moveit_cpp/launch/moveit_cpp_tutorial.launch.py

@@ -11,6 +11,7 @@ def generate_launch_description():
         MoveItConfigsBuilder("moveit_resources_panda")
         .robot_description(file_path="config/panda.urdf.xacro")
         .trajectory_execution(file_path="config/gripper_moveit_controllers.yaml")
+        .planning_pipelines("stomp", ["stomp"])
         .moveit_cpp(
             file_path=get_package_share_directory("moveit2_tutorials")
             + "/config/moveit_cpp.yaml"

doc/examples/planning_adapters/planning_adapters_tutorial.rst

@@ -6,7 +6,7 @@
 Planning Adapter Tutorials
 ==========================
 
-Planning Request Adapters is a concept in MoveIt which can be used to modify the trajectory (pre-processing and/or post-processing) for a motion planner. Some examples of existing planning adapters in MoveIt include AddTimeParameterization, FixWorkspaceBounds, FixStartBounds, FixStartStateCollision, FixStartStatePathConstraints, CHOMPOptimizerAdapter, etc. ! Using the concepts of Planning Adapters, multiple motion planning algorithms can be used in a pipeline to produce robust motion plans. For example, a sample pipeline of motion plans might include an initial plan produced by OMPL which can then be optimized by CHOMP to produce a motion plan which would likely be better than a path produced by OMPL or CHOMP alone. Similarly, using the concept of Planning Adapters, other motion planners can be mixed and matched depending on the environment the robot is operating in. This section provides a step by step tutorial on using a mix and match of different motion planners and also provides insights on when to use which particular motion planners.
+Planning Request Adapters is a concept in MoveIt which can be used to modify the trajectory (pre-processing and/or post-processing) for a motion planner. Some examples of existing planning adapters in MoveIt include AddTimeParameterization, ValidateWorkspaceBounds, FixStartBounds, CheckStartStateCollision, CHOMPOptimizerAdapter, etc. ! Using the concepts of Planning Adapters, multiple motion planning algorithms can be used in a pipeline to produce robust motion plans. For example, a sample pipeline of motion plans might include an initial plan produced by OMPL which can then be optimized by CHOMP to produce a motion plan which would likely be better than a path produced by OMPL or CHOMP alone. Similarly, using the concept of Planning Adapters, other motion planners can be mixed and matched depending on the environment the robot is operating in. This section provides a step by step tutorial on using a mix and match of different motion planners and also provides insights on when to use which particular motion planners.
 
 Getting Started
 ---------------
@@ -46,11 +46,10 @@ To achieve this, follow the steps:
 #. Open the ``ompl_planning_pipeline.launch`` file in the ``<robot_moveit_config>/launch`` folder of your robot. For the Panda robot it is this `file <https://github.com/ros-planning/panda_moveit_config/blob/melodic-devel/launch/ompl_planning_pipeline.launch.xml>`_. Edit this launch file, find the lines where ``<arg name="planning_adapters">`` is mentioned and change it to: ::
 
     <arg name="planning_adapters"
-         value="default_planner_request_adapters/AddTimeParameterization
-                default_planner_request_adapters/FixWorkspaceBounds
-                default_planner_request_adapters/FixStartStateBounds
-                default_planner_request_adapters/FixStartStateCollision
-                default_planner_request_adapters/FixStartStatePathConstraints
+         value="default_planning_request_adapters/AddTimeParameterization
+                default_planning_request_adapters/ValidateWorkspaceBounds
+                default_planning_request_adapters/CheckStartStateBounds
+                default_planning_request_adapters/CheckStartStateCollision
                 chomp/OptimizerAdapter" />
 
 #. The values of the ``planning_adapters`` is the order in which the mentioned adapters are called / invoked. Order here matters. Inside the CHOMP adapter, a :moveit_codedir:`call <moveit_planners/chomp/chomp_optimizer_adapter/src/chomp_optimizer_adapter.cpp#L169>` to OMPL is made before invoking the CHOMP optimization solver, so CHOMP takes the initial path computed by OMPL as the starting point to further optimize it.
@@ -81,11 +80,10 @@ To achieve this, follow the steps:
 
 #. Open the ``stomp_planning_pipeline.launch`` file in the ``<robot_moveit_config>/launch`` folder of your robot. For the Panda robot it is `this <https://github.com/ros-planning/panda_moveit_config/blob/melodic-devel/launch/stomp_planning_pipeline.launch.xml>`_ file. Edit this launch file, find the lines where ``<arg name="planning_adapters">`` is mentioned and change it to: ::
 
-    <arg name="planning_adapters" value="default_planner_request_adapters/AddTimeParameterization
-                   default_planner_request_adapters/FixWorkspaceBounds
-                   default_planner_request_adapters/FixStartStateBounds
-                   default_planner_request_adapters/FixStartStateCollision
-                   default_planner_request_adapters/FixStartStatePathConstraints
+    <arg name="planning_adapters" value="default_planning_request_adapters/AddTimeParameterization
+                   default_planning_request_adapters/ValidateWorkspaceBounds
+                   default_planning_request_adapters/CheckStartStateBounds
+                   default_planning_request_adapters/CheckStartStateCollision
                    chomp/OptimizerAdapter" />
 
 #. The values of the ``planning_adapters`` is the order in which the mentioned adapters are called / invoked. Order here matters. Inside the CHOMP adapter, a call to STOMP is made before invoking the CHOMP optimization solver, so CHOMP takes the initial path computed by STOMP as the starting point to further optimize it.
@@ -118,11 +116,10 @@ To achieve this, follow the steps:
 
 #. Open the ``ompl_planning_pipeline.launch`` file in the ``<robot_moveit_config>/launch`` folder of your robot. For the Panda robot it is this `file <https://github.com/ros-planning/panda_moveit_config/blob/melodic-devel/launch/ompl_planning_pipeline.launch.xml>`_. Edit this launch file, find the lines where ``<arg name="planning_adapters">`` is mentioned and change it to: ::
 
-    <arg name="planning_adapters" value="default_planner_request_adapters/AddTimeParameterization
-                   default_planner_request_adapters/FixWorkspaceBounds
-                   default_planner_request_adapters/FixStartStateBounds
-                   default_planner_request_adapters/FixStartStateCollision
-                   default_planner_request_adapters/FixStartStatePathConstraints
+    <arg name="planning_adapters" value="default_planning_request_adapters/AddTimeParameterization
+                   default_planning_request_adapters/ValidateWorkspaceBounds
+                   default_planning_request_adapters/CheckStartStateBounds
+                   default_planning_request_adapters/CheckStartStateCollision
                    stomp_moveit/StompSmoothingAdapter" />
 
 #. The values of the ``planning_adapters`` is the order in which the mentioned adapters are called / invoked. Order here matters. Inside the STOMP adapter, a call to OMPL is made before invoking the STOMP smoothing solver, so STOMP takes the initial path computed by OMPL as the starting point to further optimize it.

doc/examples/subframes/subframes_tutorial.rst

@@ -69,5 +69,5 @@ For older moveit_config packages that you have not generated yourself recently,
 required for subframes might not be configured, and the subframe link might not be found. To fix this for your
 moveit_config package, open the ``ompl_planning_pipeline.launch`` file in the ``<robot_moveit_config>/launch``
 folder of your robot. For the Panda robot it is :panda_codedir:`this <launch/ompl_planning_pipeline.launch.xml>` file.
-Edit this launch file, find the lines where ``<arg name="planning_adapters">`` is mentioned and insert ``default_planner_request_adapters/ResolveConstraintFrames`` after
-the line ``default_planner_request_adapters/FixStartStatePathConstraints``.
+Edit this launch file, find the lines where ``<arg name="planning_adapters">`` is mentioned and insert ``default_planning_request_adapters/ResolveConstraintFrames`` after
+the line ``default_planning_request_adapters/FixStartStatePathConstraints``.

doc/examples/time_parameterization/time_parameterization_tutorial.rst

@@ -34,7 +34,7 @@ Finally, add the Ruckig smoothing algorithm to the list of planning request adap
 
 .. code-block:: yaml
 
-      request_adapters: >-
-        default_planner_request_adapters/AddRuckigTrajectorySmoothing
-        default_planner_request_adapters/AddTimeOptimalParameterization
+      response_adapters:
+        - default_planning_request_adapters/AddRuckigTrajectorySmoothing
+        - default_planning_request_adapters/AddTimeOptimalParameterization
         ...

doc/how_to_guides/benchmarking/launch/run_benchmarks.launch.py

@@ -46,15 +46,20 @@ def generate_launch_description():
     # Load additional OMPL pipeline
     ompl_planning_pipeline_config = {
         "ompl_rrtc": {
-            "planning_plugin": "ompl_interface/OMPLPlanner",
-            "request_adapters": """\
-                default_planner_request_adapters/AddTimeOptimalParameterization \
-                default_planner_request_adapters/FixWorkspaceBounds \
-                default_planner_request_adapters/FixStartStateBounds \
-                default_planner_request_adapters/FixStartStateCollision \
-                default_planner_request_adapters/FixStartStatePathConstraints \
-              """,
-            "start_state_max_bounds_error": 0.1,
+            "planning_plugins": [
+                "ompl_interface/OMPLPlanner",
+            ],
+            "request_adapters": [
+                "default_planning_request_adapters/ResolveConstraintFrames",
+                "default_planning_request_adapters/ValidateWorkspaceBounds",
+                "default_planning_request_adapters/CheckStartStateBounds",
+                "default_planning_request_adapters/CheckStartStateCollision",
+            ],
+            "response_adapters": [
+                "default_planning_response_adapters/AddTimeOptimalParameterization",
+                "default_planning_response_adapters/ValidateSolution",
+                "default_planning_response_adapters/DisplayMotionPath",
+            ],
         }
     }
     ompl_planning_yaml = load_yaml(

doc/how_to_guides/chomp_planner/chomp_planner_tutorial.rst

@@ -98,31 +98,3 @@ OMPL is a open source library for sampling based / randomized motion planning al
 CHOMP: While most high-dimensional motion planners separate trajectory generation into distinct planning and optimization stages, CHOMP capitalizes on covariant gradient and functional gradient approaches to the optimization stage to design a motion planning algorithm based entirely on trajectory optimization. Given an infeasible naive trajectory, CHOMP reacts to the surrounding environment to quickly pull the trajectory out of collision while simultaneously optimizing dynamic quantities such as joint velocities and accelerations. It rapidly converges to a smooth, collision-free trajectory that can be executed efficiently on the robot. A covariant update rule ensures that CHOMP quickly converges to a locally optimal trajectory.
 
 For scenes containing obstacles, CHOMP often generates paths which do not prefer smooth trajectories by addition of some noise (*ridge_factor*) in the cost function for the dynamic quantities of the robot (like acceleration, velocity). CHOMP is able to avoid obstacles in most cases, but it can fail if it gets stuck in local minima due to a bad initial guess for the trajectory. OMPL can be used to generate collision-free seed trajectories for CHOMP to mitigate this issue.
-
-Using CHOMP as a post-processor for OMPL
-----------------------------------------
-Here, we will demonstrate that CHOMP can also be used as a post-processing optimization technique for plans obtained by other planning algorithms. The intuition behind this is that some randomized planning algorithm produces an initial guess for CHOMP. CHOMP then takes this initial guess and further optimizes the trajectory.
-To achieve this, use the following steps:
-
-#. Edit ``ompl_planning.yaml`` in the ``<robot_moveit_config>/config`` folder of your robot. Add ``chomp/OptimizerAdapter`` to the bottom of the list of request_adapters: ::
-
-    request_adapters: >-
-      ...
-      default_planner_request_adapters/FixStartStatePathConstraints
-      chomp/OptimizerAdapter
-
-#. Change the ``trajectory_initialization_method`` parameter in ``chomp_planning.yaml`` to ``fillTrajectory`` so that OMPL can provide the input for the CHOMP algorithm: ::
-
-    trajectory_initialization_method: "fillTrajectory"
-
-#. Add the CHOMP config file to the launch file of your robot, ``<robot_moveit_config>/launch/chomp_demo.launch.py``, if it is not there already: ::
-
-    .planning_pipelines(pipelines=["ompl", "chomp"])
-
-#. Now you can launch the newly configured planning pipeline as follows: ::
-
-    ros2 launch moveit2_tutorials chomp_demo.launch.py rviz_tutorial:=True
-
-This will launch RViz. Select OMPL in the Motion Planning panel under the Context tab. Set the desired start and goal states by moving the end-effector around in the same way as was done for CHOMP above. Finally click on the Plan button to start planning. The planner will now first run OMPL, then run CHOMP on OMPL's output to produce an optimized path. To make the planner's task more challenging, add obstacles to the scene using: ::
-
-    ros2 run moveit2_tutorials collision_scene_example