Research report · Active
Adaptive Multi-Modal SLAM with Uncertainty-Aware Sensor Fusion
Robust perception research · Active · 2026–Present
This project studies whether SLAM systems can estimate sensor reliability online and adapt their fusion strategy before visual degradation causes severe localization failure.
Abstract
Investigate adaptive fusion of visual, inertial, and event-based sensing under perceptual degradation such as blur, low light, texture scarcity, and rapid motion.
Scientific question: Can SLAM systems estimate sensor reliability online and adapt fusion weights before severe localization failure?
Contribution: A research scaffold for degradation-aware SLAM benchmarking and adaptive sensor fusion.
Marked research artifact slot
Adaptive fusion architecture
Architecture figure slot for verified diagrams of modules, data flow, and software boundaries.
Marked research artifact slot
Input degradation · fusion · trajectory evaluation
Pipeline slot for GIFs or animations that explain estimation, planning, perception, or evaluation steps.
Marked research artifact slot
Synthetic sensor reliability visualization
Demo slot reserved for real project videos or synthetic demos that are explicitly labeled as synthetic.
Research questions
- Can a SLAM system estimate sensor reliability online?
- When should a robot reduce reliance on visual measurements and increase reliance on inertial or event-based sensing?
- How should robustness be evaluated under controlled perceptual degradation?
Methods
Implementation status
Implemented
- Project architecture
- Evaluation scaffold
- Research framing
- Baseline integration plan
Research prototype
- Adaptive fusion logic
- Trajectory matching utilities
- Degradation scenario definitions
Planned
- Real ORB-SLAM3 baseline runs
- Event-camera experiments
- Benchmark report
Experiments and metrics
Experiments
- EuRoC parsing
- Trajectory matching
- Adaptive fusion prototype
- Baseline SLAM evaluation scaffold
Metrics
- ATE
- RPE
- sensor reliability score
- failure rate under degradation — planned
Quantitative benchmark tables will be added only after reproducible experiments are available.
Limitations
- Quantitative claims are intentionally withheld until reproducible benchmark runs are available.
- Visual figures and GIFs are placeholders unless a project page explicitly states that an artifact is generated from real experiments.
Reproducibility plan
Experiment workflow
Each module is prepared to document configuration files, datasets, evaluation metrics, and repeated experiment runs before quantitative claims are shown.
Repository setup
Public repositories are linked when available. Private, planned, or incomplete repositories remain clearly labelled rather than being presented as finished systems.
Literature context
- ORB-SLAM3
- VINS-Fusion
- OpenVINS
- Event-based vision
- Uncertainty calibration
Roadmap
- Complete real ORB-SLAM3 baseline runs
- Add event-camera degradation experiments
- Compare fixed and adaptive fusion strategies