Overview
This project focuses on developing hierarchical representations and real-time systems for spatial perception in robotics. The work addresses the fundamental challenge of enabling robots to understand and navigate complex 3D environments efficiently.
Research Goals
- Develop hierarchical spatial representations that capture both local and global structure
- Create real-time algorithms for spatial understanding
- Enable robust navigation in complex environments
- Bridge the gap between perception and planning
Key Contributions
Hierarchical Representations
- Multi-scale spatial representations
- Efficient data structures for real-time processing
- Integration of geometric and semantic information
Real-time Systems
- Optimized algorithms for real-time performance
- Hardware-accelerated implementations
- Scalable architectures for large environments
Publications
- “Foundations of Spatial Perception for Robotics: Hierarchical Representations and Real-time Systems” - Int. J Robotic Research
- “Certifiable 3D Object Pose Estimation: Foundations, Learning Models, and Self-Training” - IROS 2023
Applications
- Autonomous navigation
- Robot manipulation
- Environmental mapping
- Search and rescue operations
Collaborators
- Prof. Luca Carlone (MIT SPARKlab)
- Various researchers in robotics and computer vision
Links
