myo-model 0.0.0

myo_model is a library of carefully constructed musculoskeletal models.

MyoModel

MyoModel is a library of carefully constructed Musculoskeletal Models.

Overview

Musculoskeletal models are one the fundamental building blocks in diverse fields - biomechanics, graphics, animation, rehabilitation, etc. Given their significance, there is a rich history of musculoskeletal modeling efforts by multiple groups over decades. In addition to varying conventions, such efforts face two key challenges -

1. sparsity of experimental data - leading to localized incomplete models
2. computational challenges - in capturing the full human anatomy

The goal of MyoModel is to develop a comprehensive library that unifies the fragmented developments from disjoint fields while meeting the computational challenges involved in capturing full details of human anatomy.

Usage

There are multiple ways to explore and leverage MyoModel

• MyoModel: For any needs involving only access to the models, myo_model (this repo) is all you need
• MyoTools: For programmatic access to the models and basic developmental needs, please refer to myo_tools repo
• MyoSuite: For building data-driven behavioral controllers for the models, please refer to myosuite repo

License

A permissive license for non-commercial scientific research is available here.

Citation

@techreport{myoskeleton,
  author      = {Vittorio Caggiano AND Vittorio La Barbera AND Andrea Prestia AND Ouassim Aouattah AND Pierre Schumacher AND Varun Joshi AND Vikash Kumar},
  title       = {MyoSkeleton: A Universal Human Skeletal Model},
  institution = {MyoLab Inc.},
  year        = {2024},
  type        = {White Paper},
  note        = {Available at: \url{https://github.com/myolab/myo_model}},
}

Example usage

import mujoco
import mujoco.viewer
from myo_model.utils import model_utils

model_path = model_utils.get_model_xml_path()
mj_model = mujoco.MjModel.from_xml_path(model_path)

# Make the data object
mj_data = mujoco.MjData(mj_model)

# Run the viewer
with mujoco.viewer.launch_passive(
    mj_model, mj_data) as viewer:
    while viewer.is_running():
        viewer.sync()