Skip to content

How to use the new joint-space planner introduced in cuMotion 4.0 #44

@Namsangee

Description

@Namsangee

Description

Hello NVIDIA Isaac ROS team,

According to the latest changelog for Isaac ROS cuMotion 4.0, the update on 2025-10-24 mentions:

“Added joint space planner and support for PyNITROS.”

I would like to confirm how the newly added joint-space planner can be enabled and used within the current MoveIt 2 + cuMotion integration.

From reviewing the latest source code of CumotionActionServer, it still appears that:

The planning request (MotionGen.plan_single) uses a Pose (task-space) goal.

Joint-based goals (goal_constraints.joint_constraints) are internally converted to an end-effector pose through forward kinematics before planning.

It is unclear whether the new joint-space planner is accessible via a planner ID, configuration parameter, or a different API call.

Questions

How can we explicitly activate the joint-space planner introduced in the 2025-10-24 update?

Is it exposed via a MoveIt 2 planner ID such as isaac_ros_cumotion_joint or through a parameter inside isaac_ros_cumotion_planning.yaml?

Does this planner perform true joint-space optimization (for example, minimizing joint distances or velocity cost) rather than pose error minimization?

Can it be used interchangeably with the default Cartesian cuMotion planner inside the same MoveIt 2 configuration?

Are there example configurations or launch files demonstrating joint-space planner usage?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions