Robodimm

Robot Actuator Sizing Made Simple

Design, simulate, and optimize industrial robot motions.
Choose your mode to get started.

<sub>Developed by

Custom Robotics</sub>

---

Robodimm is a web application for robot motion programming, inverse dynamics, and automated actuator sizing for scalable industrial robots. It combines a physics-based backend with an interactive 3D frontend to provide an end-to-end toolchain: from trajectory specification to motor and gearbox selection.

Key features:

• DEMO mode — runs entirely in the browser, no server required.
• PRO mode — full physics backend via FastAPI + Pinocchio + Pink.
• CR4 4-DOF palletizer (parallelogram closed-loop) and CR6 6-DOF spherical-wrist robot models.
• Constrained inverse dynamics (KKT), actuator requirement analysis, and motor/gearbox selection.
• Geometric scaling laws for structural mass/inertia with configurable exponents.

Live deployment

Try it now at https://customrobotics.es/

Interface

DEMO mode — browser-only simulation: joint jogging, target programming, and trapezoidal motion profiles:

PRO mode demo — real-time inverse dynamics, IK solver, and actuator sizing (click to play):

demo_robodimm.mp4

Quick Start

Robodimm has two operating modes with different requirements:

	DEMO mode	PRO mode
Physics backend (Pinocchio, Pink)	❌ browser only	✅ required
Inverse dynamics (KKT)	❌	✅
IK solver (Pink QP)	❌	✅
REST API (FastAPI)	❌	✅
Installation required	none	Docker / Conda

DEMO mode — no installation required

All logic runs in the browser (JavaScript). No server or Python dependencies needed.

python3 -m http.server 8080 --directory frontend

Open http://localhost:8080/simulator.html?mode=demo

PRO mode — full physics backend

Docker (recommended) — single command, no dependency management:

cp .env.local .env
docker compose up --build

Open http://localhost:8000/ or http://localhost:8000/simulator?mode=pro

Conda — for development:

conda env create -f environment.yml
conda activate robodimm_env
uvicorn backend.main:app --reload --host 0.0.0.0 --port 8000

pip — minimal setup:

pip install -r requirements.txt
uvicorn backend.main:app --reload --host 0.0.0.0 --port 8000

Default login: admin / robotics

Running the Tests

The test suite requires Pinocchio and Pink (included in environment.yml and the Docker image).

Docker (recommended):

docker compose build
docker compose run --rm robodimm micromamba run -n base pytest tests/ -v

Conda:

conda activate robodimm_env
pytest tests/ -v

The suite contains 214 unit tests covering:

Module	Tests	Coverage
robot_core.interpolation	22	Trapezoidal/linear profiles, multi-DOF trajectories
robot_core.inertial_params	25	Box/cylinder inertia, CR4 scaling laws
robot_core.actuators	23	Trajectory requirements, motor/gearbox selection
robot_core.kinematics	22	FK (forward kinematics), IK solver, build_robot
robot_core.dynamics	15	KKT constrained ID, trajectory dynamics, method comparison
robot_core conversions	16	Real ↔ Pink ↔ RobotStudio joint space round-trips
backend.models	43	Pydantic schema validation for all API endpoints
backend.utils	22	Coordinate conversions, demo targets, index mapping

All tests pass in < 1 second on the Docker image (Python 3.10, Pinocchio 3.x, Pink 4.x).

Project Layout

robodimm-main/
├── backend/          # FastAPI app and API routers
├── frontend/         # Web UI and DEMO-side logic (Three.js)
├── robot_core/       # Robot builders, kinematics, dynamics, actuators
│   ├── builders/     # CR4 and CR6 Pinocchio model builders
│   └── dynamics/     # Constrained ID, trajectory dynamics
├── tests/            # 214 unit tests (pytest)
├── meshes/           # Robot mesh assets (GLTF/GLB)
├── docs/             # Documentation and images
├── environment.yml   # Conda environment (Python 3.10 + Pinocchio + Pink)
├── Dockerfile        # Micromamba-based container
└── actuators_library.json  # Shared actuator catalog

Main Capabilities

• Joint and Cartesian jog with real-time IK
• Target and program editing with MoveJ / MoveL / MoveC instructions
• DEMO/PRO execution behind a common frontend workflow
• Trajectory inverse dynamics (RNEA + KKT constraints) and CSV export
• Actuator library management and motor/gearbox selection with safety factors
• Geometric scaling of structural mass and inertia (configurable exponents)
• Payload, friction, reflected inertia, and motor stator mass configuration

Notes

• PRO mode is expected on port 8000.
• DEMO clears local storage on load unless ?keep_saved=true is used.
• The shared actuator library uses anonymized generic labels such as AC_* and HD*.
• Structural scaling parameters are configurable in the app; the paper case-study values are not hardcoded defaults.

Additional Docs

• QUICK_START.md
• DEPLOYMENT.md

Publication

Robodimm is described in the following paper:

J. L. Torres, M. Munoz, J. D. Alvarez, J. L. Blanco, and A. Gimenez, "Robodimm: A Physics-Grounded Framework for Automated Actuator Sizing in Scalable Modular Robots," arXiv:2603.06864, 2026.

• arXiv: https://arxiv.org/abs/2603.06864
• DOI: https://doi.org/10.48550/arXiv.2603.06864

How to Cite

@article{torres2026robodimm,
  title   = {Robodimm: A Physics-Grounded Framework for Automated Actuator Sizing in Scalable Modular Robots},
  author  = {Torres, J. L. and Munoz, M. and Alvarez, J. D. and Blanco, J. L. and Gimenez, A.},
  journal = {arXiv preprint arXiv:2603.06864},
  year    = {2026},
  doi     = {10.48550/arXiv.2603.06864},
  url     = {https://arxiv.org/abs/2603.06864}
}

License

This project is licensed under the MIT License.