[Bug Report] Unable to run example add a new robot to Isaac Lab

[theunknowninfinite]

If you are submitting a bug report, please fill in the following details and use the tag [bug].

Describe the bug

Unable to run example using code from https://github.com/isaac-sim/IsaacLab/blob/develop/scripts/tutorials/01_assets/add_new_robot.py.

Steps to reproduce

• Running code using python3 scripts/tutorials/01_assets/add_new_robot.py
• Using ./isaaclab.sh -p scripts/tutorials/01_assets/add_new_robot.py
  Terminal logs:

/home/infinity/lab/scripts/tutorials/01_assets/add_new_robot.py:151: DeprecationWarning: The function 'set_joint_velocity_target' will be deprecated in a future release. Please use 'set_joint_velocity_target_index' instead.
  scene["Jetbot"].set_joint_velocity_target(action)
Traceback (most recent call last):
  File "/home/infinity/lab/scripts/tutorials/01_assets/add_new_robot.py", line 183, in <module>
    main()
  File "/home/infinity/lab/scripts/tutorials/01_assets/add_new_robot.py", line 179, in main
    run_simulator(sim, scene)
  File "/home/infinity/lab/scripts/tutorials/01_assets/add_new_robot.py", line 151, in run_simulator
    scene["Jetbot"].set_joint_velocity_target(action)
  File "/home/infinity/lab/source/isaaclab/isaaclab/assets/articulation/base_articulation.py", line 2595, in set_joint_velocity_target
    self.set_joint_velocity_target_index(target=target, joint_ids=joint_ids, env_ids=env_ids)
  File "/home/infinity/lab/source/isaaclab_physx/isaaclab_physx/assets/articulation/articulation.py", line 2334, in set_joint_velocity_target_index
    wp.launch(
  File "/home/infinity/miniconda3/envs/lab/lib/python3.12/site-packages/warp/_src/context.py", line 7468, in launch
    pack_args(fwd_args, params, adjoint=False)
  File "/home/infinity/miniconda3/envs/lab/lib/python3.12/site-packages/warp/_src/context.py", line 7428, in pack_args
    params.append(pack_arg(kernel, arg_type, arg_name, a, device, adjoint))
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/infinity/miniconda3/envs/lab/lib/python3.12/site-packages/warp/_src/context.py", line 6905, in pack_arg
    raise RuntimeError(
RuntimeError: Error launching kernel 'write_2d_data_to_buffer_with_indices', argument 'in_data' expects an array of type array(ndim=2, dtype=float32), but passed value has type Tensor.
2026-05-03T22:46:02Z [4,534ms] [Warning] [simulation_app.simulation_app] SimulationApp.close() was not called explicitly. Shutting down automatically

-->

System Info

Describe the characteristic of your environment:

• Commit: release v3.0.0-beta
• Isaac Sim Version: 6.0.0
• OS: Linux Mint 22.3
• GPU: RTX 5090
• CUDA: 13
• GPU Driver: 580.142

Additional context

Add any other context about the problem here.

Checklist

• I have checked that there is no similar issue in the repo (required)
• I have checked that the issue is not in running Isaac Sim itself and is related to the repo

Acceptance Criteria

Add the criteria for which this task is considered done. If not known at issue creation time, you can add this once the issue is assigned.

• Criteria 1
• Criteria 2

[RandomOakForest]

Thank you for posting this. The team will review it and follow up.

[Exokern]

This looks like a type mismatch between the tutorial and the PhysX backend:

• The tutorial scripts/tutorials/01_assets/add_new_robot.py builds action = torch.Tensor([[10.0, 10.0]]) (a torch.Tensor).
• BaseArticulation.set_joint_velocity_target() forwards that into set_joint_velocity_target_index(), which ends up calling wp.launch(...) in isaaclab_physx/assets/articulation/articulation.py.
• Warp expects a wp.array there; the error you see (expects array(ndim=2, dtype=float32), but passed value has type Tensor) is consistent with passing a torch tensor straight into a Warp kernel.

Quick workaround in the tutorial: make action a wp.array (float32, correct device) before calling set_joint_velocity_target[_index].

Upstream fix options (small):

1. Fix the tutorial to pass a wp.array.
2. Or make set_joint_velocity_target_index() accept torch.Tensor and internally convert to a wp.array before the wp.launch(...) call (so the public API can consistently accept torch inputs).

Happy to send a tiny PR for either approach if maintainers have a preference.

[RandomOakForest]

Following up, the team is actively working to resolve this. @kellyguo11 for vis.

[lqxisok]

The follow code works well for me:

  # drive around
  if count % 100 < 75:
      # Drive straight by setting equal wheel velocities
      action = wp.from_torch(torch.Tensor([[10.0, 10.0]]).cuda())
  else:
      # Turn by applying different velocities
      action = wp.from_torch(torch.Tensor([[5.0, -5.0]]).cuda())