[Bug Report] Bugs of body_index & skrl in Orbit

[ZeyuGao-2022]

Describe the bug

Hello, I've found two bugs in Orbit. I'm not sure that description is certainty, but I hope this Bug Report can help you fix these bugs in the next version.

1. Bug of body_index The function _process_info_cfg which define self.ee_body_index = body_index in "source/extensions/omni.isaac.orbit/omni/isaac/orbit/robots/single_arm.py", line 156. The ee_body_index should minus 1. The effort of this var is find the ee_body's Jacobian. For example, my baxter robot has 57 links, the index of base is 0. However, the first dim of Jocabian is 56 except the base. The ee_body_index should minus 1 to Ensure one-to-one correspondence between Jacobian and index.

   def _process_info_cfg(self) -> None:
       """Post processing of configuration parameters."""
       # process parent config
       super()._process_info_cfg()
       # resolve regex expressions for indices
       # -- end-effector body
       self.ee_body_index = None
       for body_index, body_name in enumerate(self.body_names):
           if re.fullmatch(self.cfg.ee_info.body_name, body_name):
               self.ee_body_index = body_index - 1

2. Bug of skrl This is a simple bug which was founded when I was running the demo of skrl. But I'm not sure is the version of environment causes this bug.

   Traceback (most recent call last):
   File "source/standalone/workflows/skrl/train.py", line 167, in <module>
   main()
   File "source/standalone/workflows/skrl/train.py", line 107, in main
   **convert_skrl_cfg(experiment_cfg["models"]["policy"]),
   TypeError: gaussian_model() argument after ** must be a mapping, not NoneType

I fix it by adding "return d" after "source/standalone/workflows/skrl/config.py", line 90.

def update_dict(d):
        for key, value in d.items():
            if isinstance(value, dict):
                update_dict(value)
            else:
                if key in _direct_eval:
                    d[key] = eval(value)
                elif key.endswith("_kwargs"):
                    d[key] = value if value is not None else {}
                elif key in ["rewards_shaper_scale"]:
                    d["rewards_shaper"] = reward_shaper_function(value)
        return d

System Info

Describe the characteristic of your environment:

• Commit: db71160
• Isaac Sim Version: 2022.2.0
• OS: Ubuntu 20.04
• GPU: RTX 4090
• CUDA: 12.0
• GPU Driver: 525.85.05

Checklist

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

[Mayankm96]

Thanks a lot for pointing these out! I made a PR #22 to fix the issues.

Instead of doing self.ee_body_index = body_index - 1, I feel it is better to account for the subtraction inside the Jacobian function. It prevents confusion if someone externally tries to access ee_body_index for some other reasons.

Once you approve the MR, I can merge it.