Releases: Synria-Robotics/RoboCore
Releases · Synria-Robotics/RoboCore
RoboCore v2.0.0
✨ Key Features
| Module | Functionality | Status |
|---|---|---|
| Modeling | URDF/MJCF parsing, Robot abstraction | ✅ Stable |
| Kinematics | FK/IK (DLS/Pinv/Transpose), Batch processing | ✅ Stable |
| Bimanual FK/IK (indep/relative/mirror modes) | ✅ Stable | |
| Jacobian | Analytic/Numeric/Autograd methods | ✅ Stable |
| Transform | SE(3)/SO(3) operations | ✅ Stable |
| Planning | Trajectory generation | ✅ Stable |
| Control | Joint/Cartesian controllers | ✅ Stable |
| Analysis | Workspace/Singularity analysis | ✅ Stable |
| WDF | SDF/RDF (Distance Fields) | ✅ Stable |
| Bridge | MuJoCo simulation bridge | ✅ Stable |
🚀 Performance Benchmarks
Test Platform: Intel i7-10700K, NVIDIA RTX 3080
CPU (NumPy): Up to 107x faster Jacobian computation and 80x faster IK compared to pure Python.
GPU (PyTorch): 14x speedup on batch Forward Kinematics (1000 configs) vs CPU.
| Operation | Pure Python | NumPy | Speedup |
|---|---|---|---|
| Forward Kinematics | 2.5 ms | 0.05 ms | 50x |
| Inverse Kinematics | 450 ms | 5.6 ms | 80x |
| Jacobian (Analytic) | 3.2 ms | 0.03 ms | 107x |
| Operation | NumPy (CPU) | PyTorch (GPU) | Speedup |
|---|---|---|---|
| Forward Kinematics (1000) | 45 ms | 3.2 ms | 14x |
| Jacobian (1000) | 28 ms | 2.1 ms | 13x |
🆕 What's New
- Bimanual Support: Dual-arm kinematics with indep/relative/mirror modes
- Control Module: Joint/Cartesian position/velocity/trajectory controllers
- WDF: Signed/Relative Distance Fields for workspace analysis
- MuJoCo Bridge: Physics simulation and trajectory evaluation
- Multi-chain Support: Complex robot topologies
📦 Installation
pip install synria-robocoreRoboCore v1.0.0
v1.0.0
✨ Key Features
| Module | Functionality | Status |
|---|---|---|
| Modeling | URDF/MJCF parsing, Robot abstraction | ✅ Stable |
| Kinematics | FK/IK (DLS/Pinv), Batch processing | ✅ Stable |
| Jacobian | Analytic/Numeric/Autograd methods | ✅ Stable |
| Transform | SE(3)/SO(3) operations | ✅ Stable |
| Planning | Trajectory generation | 🚧 Alpha |
🚀 Performance Benchmarks
RoboCore is optimized for speed. Benchmarks performed on an Intel i7-10700K + NVIDIA RTX 3080:
- Cpu (NumPy): Up to 107x faster Jacobian computation and 80x faster IK compared to pure Python.
- GPU (PyTorch): 14x speedup on batch Forward Kinematics (1000 configs) vs CPU.