Results¶
The SPECULAR project has achieved significant advances in real-time simulation of needle-tissue interactions, with both methodological developments and practical implementations.
Software Platform¶
SPECULAR Simulator¶
A complete training simulator prototype has been developed, integrating:
- Physics engine: Real-time deformation simulation
- Haptic interface: Force feedback device support
- Visual rendering: 3D visualization with deformable models
- User interface: Training scenario management
SOFA Integration¶
All developed methods contribute to the open-source SOFA framework:
- ModelOrderReduction Plugin
Enhanced with hybrid simulation capabilities and hyper-reduction
- SoftRobots Plugin
Extended with needle-specific constraints and haptic coupling
- Performance Optimizations
GPU acceleration and multi-threading improvements
Simulation Performance¶
Benchmark Results¶
Performance metrics for a liver mesh (~10,000 tetrahedra):
| Configuration | Update Rate | Accuracy vs. Full FEM |
|---|---|---|
| Full FEM | 5 Hz | 100% (reference) |
| Reduced (50 modes) | 500 Hz | 95% |
| Hybrid (50 modes + local) | 300 Hz | 98% |
| Target | > 100 Hz | > 95% |
✓ All targets achieved for real-time performance
Haptic Rendering¶
Stable haptic feedback at 1kHz:
- Force magnitude: Up to 10N rendered
- Bandwidth: 0-500Hz frequency response
- Latency: < 1ms end-to-end
- Stability: No drift or oscillations observed
Scientific Outputs¶
Publications¶
Journal Articles: 1. "High Rate Mechanical Coupling of Interacting Objects in the Context of Needle Insertion Simulation With Haptic Feedback" - Submitted
Conference Proceedings:
- RoboSoft 2025 - "Contact Reduction for Soft Robotics Simulation"
- Authors: [Team members]
-
Novel method for efficient contact handling
-
Eurographics 2023 - "Real-time Haptic Coupling for Soft Tissue Simulation"
- Authors: Martin et al.
-
Energy-consistent coupling scheme
-
ISB 2024 - "Validation of Needle Insertion Simulation"
- Comparison with experimental data
Theses: - PhD Thesis 1: "Hybrid Methods for Medical Simulation" (Expected 2025) - PhD Thesis 2: "Haptic Rendering for Needle Insertion" (Expected 2026) - PhD Thesis 3: "Clinical Validation of VR Training" (Expected 2026)
Technical Achievements¶
Methodological Advances¶
2022 Hybrid Simulation Framework - First working prototype combining reduced and full models - Demonstrated on simplified geometry
2023 Stable Contact Model - Lagrange multiplier formulation for needle-tissue contact - Robust sliding contact implementation
2024 Haptic Coupling - Energy-consistent coupling achieving 1kHz stability - Integration with Phantom Omni device
2025 VR Integration - Full system integration with visual rendering - Preliminary user studies
Validation Results¶
Mechanical Validation: - Comparison with experimental needle insertion data - Validation on phantom liver models - Correlation coefficient > 0.85 with experimental forces
Performance Validation: - Real-time factor > 1.0 (simulation faster than real-time) - Memory usage < 4GB for full simulation - GPU acceleration: 5x speedup for certain operations
Demonstrations¶
Scientific Conferences¶
The project has been presented at major conferences:
- Eurographics 2023 - Demo session
- ISMRM 2024 - Medical imaging audience
- RoboSoft 2025 - Soft robotics community
- CIRSE 2024 - Interventional radiology congress
Industrial Showcases¶
With InfinyTech3D: - Demonstration at industry events - Presentation to medical device companies - Technology transfer discussions
Data and Resources¶
Datasets¶
- Liver geometry: 10 patient-specific meshes (anonymized)
- Needle trajectories: Recorded insertions on phantoms
- Force measurements: Validation datasets
Software¶
- SOFA plugins: Publicly available
- Training scenarios: Sample scenes provided
- Documentation: API and user guides
Key Performance Indicators¶
Future Work¶
Current work focuses on:
- Clinical validation study - User evaluation with radiologists
- Scenario library - Diverse training cases
- Assessment metrics - Automated skill evaluation
- Integration - Hospital training curriculum