An open-source Python toolkit for simulating and optimizing assistive devices in neuromechanical simulations
MyoAssist is a package within MyoSuite, a collection of musculoskeletal environments built on MuJoCo for reinforcement learning and control research. It is developed and maintained by the NeuMove Lab at Northeastern University. We aim to bridge neuroscience, biomechanics, robotics, and machine learning to advance the design of assistive devices and deepen our understanding of human movement.
MyoAssist consists of three main components that together support simulation, training, and analysis of human–device interaction:
Forward simulations that combine musculoskeletal models with assistive devices.
- Currently available:
- Lower-limb exoskeletons and robotic prosthetic legs
- Planned additions:
- Upper-body wearable devices: prosthetic arms, back orthoses, etc.
- Non-wearable assistive devices: wheelchairs, externally actuated supports, etc.
- Includes baseline controllers for common assistive scenarios
Tools to generate control policies or optimize behavior in simulation.
- Reinforcement Learning (RL)
- Framework: Built on Stable-Baselines3 and PyTorch
- RL methods: Standard reinforcement learning, imitation learning, and transfer learning
- Network architecture: Modular multi-actor networks for separately controlling human and exoskeleton agents
- Controller Optimization (CO)
- Reflex-based control models
- CMA-ES for parameter tuning
A curated dataset of human movement, both real and simulated.
- Multi-Model Support: 22-muscle 2D, 26-muscle 3D, and 80-muscle musculoskeletal models
- Exoskeleton Integration: Support for multiple exoskeleton platforms (Dephy, HMEDI, Humotech, OSL, and more to come soon!)
- Reinforcement Learning: PPO-based training environments for imitation learning and terrain adaptation
- Reflex Control Optimization: CMA-ES based optimization for neuromuscular reflex controllers
- Modular Architecture: Extensible framework for custom model and controller development
- Documentation: Comprehensive guides for simulation, optimization, and analysis
- Installation
- Quick Start
- Available Models
- Project Structure
- Core Components
- Documentation
- Contributing
- License
- Related Projects
- Python 3.11+
- MuJoCo 3.3.3
- Git
-
Clone this repository:
git clone https://github.com/neumovelab/myoassist.git cd myoassist -
Set up virtual environment (recommended):
# Linux/macOS python3.11 -m venv .my_venv source .my_venv/bin/activate # Windows py -3.11 -m venv .my_venv .my_venv\Scripts\activate
-
Install the package:
pip install -e . -
Verify installation:
python test_setup.py
Please refer to the documentation for the latest Quick Start instructions:
- Getting Started: Getting Started
- Documentation Home: https://myoassist.neumove.org
| Model Type | Variant | File Name | Location | Description |
|---|---|---|---|---|
| 22-muscle 2D | BASELINE | myoLeg22_2D_BASELINE.xml |
models/22muscle_2D/ |
Basic 2D leg model without exoskeleton |
| DEPHY | myoLeg22_2D_DEPHY.xml |
models/22muscle_2D/ |
Baseline with Dephy exoskeleton | |
| HMEDI | myoLeg22_2D_HMEDI.xml |
models/22muscle_2D/ |
Baseline with HMEDI exoskeleton | |
| HUMOTECH | myoLeg22_2D_HUMOTECH.xml |
models/22muscle_2D/ |
Baseline with Humotech exoskeleton | |
| OSL_A | myoLeg22_2D_OSL_A.xml |
models/22muscle_2D/ |
Baseline with OSL ankle prosthetic | |
| TUTORIAL | myoLeg22_2D_TUTORIAL.xml |
models/22muscle_2D/ |
Tutorial model for learning purposes | |
| 26-muscle 3D | BASELINE | myoLeg26_BASELINE.xml |
models/26muscle_3D/ |
Basic 3D leg model without exoskeleton |
| DEPHY | myoLeg26_DEPHY.xml |
models/26muscle_3D/ |
3D Baseline with Dephy exoskeleton | |
| HMEDI | myoLeg26_HMEDI.xml |
models/26muscle_3D/ |
3D Baseline with HMEDI exoskeleton | |
| HUMOTECH | myoLeg26_HUMOTECH.xml |
models/26muscle_3D/ |
3D Baseline with Humotech exoskeleton | |
| OSL_A | myoLeg26_OSL_A.xml |
models/26muscle_3D/ |
3D Baseline with OSL ankle prosthetic | |
| TUTORIAL | myoLeg26_TUTORIAL.xml |
models/26muscle_3D/ |
3D tutorial model | |
| 80-muscle 3D | DEPHY | myolegs_DEPHY.xml |
models/80muscle/myoLeg80_DEPHY/ |
Full myoLegs model with Dephy exoskeleton |
| HMEDI | myolegs_HMEDI.xml |
models/80muscle/myoLeg80_HMEDI/ |
Full myoLegs model with HMEDI exoskeleton | |
| HUMOTECH | myolegs_HUMOTECH.xml |
models/80muscle/myoLeg80_HUMOTECH/ |
Full myoLegs model with Humotech exoskeleton | |
| OSL_KA | myolegs_OSL_KA.xml |
models/80muscle/myoLeg80_OSL_KA/ |
Full myoLegs model with OSL knee-ankle prosthetic |
| Asset Type | Location | Description |
|---|---|---|
| Anatomical Meshes | models/mesh/ |
Individual bone and joint STL files |
| Exoskeleton Meshes | models/mesh/Dephy/ |
Dephy exoskeleton components |
models/mesh/HMEDI/ |
HMEDI exoskeleton components | |
models/mesh/Humotech/ |
Humotech exoskeleton components | |
models/mesh/OSL/ |
OSL ankle prosthetic components | |
models/mesh/Tutorial/ |
Tutorial exoskeleton components |
myoassist/
├── Core Framework
│ ├── ctrl_optim/ # Reflex control optimization
│ ├── rl_train/ # Reinforcement learning environments
│ ├── myoassist_utils/ # Shared utilities
│ └── myosuite/ # Base musculoskeletal simulation
│
├── Models & Assets
│ ├── models/
│ │ ├── 22muscle_2D/ # 2D musculoskeletal models
│ │ ├── 26muscle_3D/ # 3D musculoskeletal models
│ │ ├── 80muscle/ # Full myoLegs models
│ │ └── mesh/ # 3D mesh assets
│ └── terrain_config.xml # Terrain configuration
│
├── Documentation
│ ├── docs/ # Framework documentation
│ └── README.md # You are here!
│
└── Configuration
├── setup.py # Package configuration
├── requirements.txt # Dependencies
└── test_setup.py # Installation verification
Reinforcement learning environments for musculoskeletal control.
Key Features:
- PPO-based training environments
- Imitation learning from reference data
- Terrain adaptation capabilities
- Multi-agent training support
Structure:
rl_train/
├── envs/ # RL environments
│ ├── myoassist_leg_base.py
│ ├── environment_handler.py
│ └── myoassist_leg_imitation.py
├── train/ # Training infrastructure
│ ├── policies/ # Policy implementations
│ ├── train_configs/ # Training configurations
│ └── train_commands/ # Training scripts
├── analyzer/ # Result analysis tools
│ ├── gait_analyze.py
│ ├── gait_evaluate.py
│ └── gait_data.py
└── reference_data/ # Training reference data
Neuromuscular reflex control optimization framework.
Key Features:
- CMA-ES optimization for controller parameters
- Multi-stage cost functions for gait objectives
- Support for 4-parameter and n-point spline controllers
- GUI and CLI tools for result analysis
Structure:
ctrl_optim/
├── train.py # Main optimization entry point
├── config/ # Configuration and argument parsing
├── cost_functions/ # Cost evaluation logic
├── optim_utils/ # CMA-ES optimizer utilities
├── preoptimized/ # Preoptimized control parameters
├── ctrl/ # Core controllers
│ ├── reflex/ # Reflex controllers
│ └── exo/ # Exoskeleton controllers
├── results/ # Optimization outputs
└── training_configs/ # Predefined configurations
Base musculoskeletal simulation framework.
Key Features:
- MuJoCo-based physics simulation
- Modular environment architecture
- Comprehensive musculoskeletal models
- Rendering and visualization tools
Structure:
myosuite/
├── envs/ # Environment implementations
├── physics/ # Physics simulation
├── renderer/ # Visualization tools
├── utils/ # Utility functions
├── simhive/ # Simulation assets
└── agents/ # Baseline controllers
- Installation Guide: Complete setup instructions
- Modeling Guide: Musculoskeletal modeling details
- RL Tutorial: Comprehensive RL guide
- Controller Optimization: Optimization
We welcome contributions!
- Please contact us for more information or if you would like to see your company's or lab's device as part of MyoAssist
- For RL questions, contact Hyoungseo Son: son.hyo@northeastern.edu
- For Reflex or modeling questions, contact Calder Robbins: robbins.cal@northeastern.edu
This project is licensed under the Apache License 2.0 - see LICENSE for details.
- MyoSuite: Base musculoskeletal simulation framework
- MuJoCo: Physics simulation engine
For questions and support, please open an issue on the project repository.