RoboCore: Unified High-Throughput Robotics Library

Developed by Synria Robotics Co., Ltd. 🤖

✨ Features

Supported Robot Formats

• ✅ URDF (Unified Robot Description Format)
• ✅ MJCF (MuJoCo XML) - Subset implementation for serial chains

Backend Support

• ✅ NumPy - CPU-optimized, 50-100x faster than pure Python
• ✅ PyTorch - GPU acceleration for batch operations

🚀 Performance Benchmarks

Test Platform: Intel i7-10700K, NVIDIA RTX 3080, 6-DOF Manipulator

Single Configuration

Batch Processing (1000 configs)

📦 Installation

# Clone repository
git clone https://github.com/Synria-Robotics/RoboCore.git
cd RoboCore

# Install (development mode)
pip install -e .

# Optional: Install with PyTorch for GPU support
pip install torch torchvision

🎯 Quick Start

Basic Example

from robocore.modeling import RobotModel

# Load robot (auto-detects URDF/MJCF)
robot = RobotModel("path/to/robot.urdf")

# Display model info
robot.summary(show_chain=True)
robot.print_tree()

# Forward Kinematics
q = [0.0] * robot.num_dof()
pose = robot.fk(q, backend='numpy', return_end=True)

# Inverse Kinematics
result = robot.ik(pose, q_initial=q, method='pinv')
print(f"Solution: {result['q']}, Success: {result['success']}")

# Jacobian
J = robot.jacobian(q, method='analytic')  # Shape: (6, dof)

Batch Processing (GPU)

import torch

# Generate random configurations
q_batch = robot.random_q_batch(batch_size=1000)

# Batch FK on GPU
poses = robot.fk(
    torch.tensor(q_batch), 
    backend='torch', 
    device='cuda',
    return_end=True
)

📚 Examples

# Robot model loading and validation
python examples/modeling/demo_robot_model.py --validate --show-tree

# Forward/Inverse kinematics
python examples/kinematics/demo_fk.py
python examples/kinematics/demo_ik.py

# Jacobian computation
python examples/kinematics/demo_jacobian.py

# Workspace analysis
python examples/analysis/demo_workspace.py --samples 10000

# Performance benchmark
python examples/kinematics/benchmark.py

🧪 Testing

# Run all tests
pytest

# Specific test suites
pytest test/unit/          # Unit tests
pytest test/integration/   # Integration tests

# With coverage report
pytest --cov=robocore --cov-report=html

📋 TODO & Roadmap

High Priority 🔴

• Collision Detection - Mesh-based collision checking
• Path Planning - RRT/RRT*/PRM algorithms
• Dynamics - Inverse/Forward dynamics computation
• Control - PID, MPC, impedance controllers

Medium Priority 🟡

• Optimization - Further SIMD/vectorization improvements
• Mobile Manipulators - Support for mobile bases
• Multi-Arm Systems - Coordinated multi-robot control

Low Priority 🟢

• Visualization - 3D interactive visualization (PyBullet/MuJoCo)
• Documentation - API docs, tutorials, best practices

🏗️ Project Structure

RoboCore/
├── robocore/
│   ├── modeling/          # Robot model abstraction & parsers
│   ├── kinematics/        # FK/IK/Jacobian solvers
│   ├── transform/         # SE(3)/SO(3) operations
│   ├── planning/          # Motion planning (WIP)
│   ├── analysis/          # Workspace/singularity analysis
│   ├── configs/           # Configuration management
│   └── utils/             # Backend abstraction, utilities
├── examples/              # Demo scripts
├── test/                  # Unit & integration tests
└── docs/                  # Documentation

📄 License

GPL-3.0 License
Copyright © 2025 Synria Robotics Co., Ltd.

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

See the LICENSE file for the full license text.

📧 Contact

• Website: synriarobotics.ai
• Email: support@synriarobotics.ai

📖 Citation

@software{robocore2025,
  title = {RoboCore: High-Performance Robotics Kinematics Library},
  author = {Synria Robotics Team},
  year = {2025},
  publisher = {Synria Robotics Co., Ltd.},
  url = {https://github.com/Synria-Robotics/RoboCore},
  version = {1.0.0}
}