search

robfiras / loco-mujoco

Imitation learning benchmark focusing on complex locomotion tasks using MuJoCo.
MIT License
475 stars 38 forks source link

the motion trajectory generated by opensim into mujoco for simulation #22

Closed 2802029572 closed 1 month ago

2802029572 commented 2 months ago

Hi, do you know how to import the motion trajectory generated by opensim into mujoco for simulation? I would be grateful if you could go into details!

robfiras commented 2 months ago

What you need is either a matlab or a numpy dataset. Both need to contain trajectory information for each joint. I would suggest you to take a look into the data generation documentation, and also the format of the matlab or numpy dataset. If you decide to go for the matlab dataset, you have to follow the data generation documentation to convert it to numpy before importing in locomujoco. Let me know if you need further help with this!

2802029572 commented 2 months ago

Thank you for your reply, but I am having problems downloading the original dataset for motion capture, is it convenient to provide the initial dataset, thanks!

robfiras commented 2 months ago

You mean you can't download them? Can you try to download them from here?

2802029572 commented 2 months ago

Thank you for your reply，I'm using from loco_mujoco.utils.dataset import download_raw_mocap_datasets download_raw_mocap_datasets() I'm having problems downloading the raw motion capture data, can I have the raw .mot files please?

robfiras commented 1 month ago

I need to know if the problem you have is actually accessing the website where the data is. Hence, could you try to download the dataset from here?

2802029572 commented 1 month ago

Hello, thank you very much for the data support, which provides and its important help to my study. If now I want to use my own human skeleton model for data replay, how should I create my own human skeleton model in mujoco? Thank you for your help!

robfiras commented 1 month ago

You will have to convert you models to mujoco models. To do so, you could use this converter. Then you create a new environment similar to the ones already created in this repo.

I am closing this, since it is not an issue.