From static 3D viewers toward operational digital representations.
Research Question
How can heterogeneous geometry, semantics, sensor context and simulation become operational digital representations rather than static 3D viewers?
Problem
The term “digital twin” is applied to everything from static 3D models to fully operational system replicas. The research challenge is bridging the gap between visualization (what something looks like) and operational intelligence (what is happening, what will happen, and what should be done).
A useful digital twin requires more than geometry: it needs semantic understanding of components, temporal state from sensors or simulation, behavioral models, and integration with operational decision-making.
Why It Matters
- Predictive maintenance and equipment lifecycle management
- Facility operations optimization
- Emergency response planning and simulation
- Urban planning and infrastructure management
- Transportation system modeling
- Industrial process monitoring
Current Approach
The research explores progressive twin maturity through multiple prototypes:
NexusTwin (Industrial)
Web-native industrial digital twin with real UCI machine-learning dataset integration, 3D visualization, AI predictive maintenance analytics, and VR readiness.
Metro Station Simulator
Agent-based passenger flow simulation with emergency evacuation modeling and real-time density heatmaps.
Stadium Digital Twin
Large-venue twin with simulated IoT sensors (temperature, occupancy, energy) and facility management dashboards.
Open London 3D
Urban-scale 3D navigation from OpenStreetMap data with camera modes and lighting simulation.
Digital Twin Maturity Classification
- 3D Viewer — Static or interactive geometry visualization
- Simulation — Contains modeled dynamic behavior
- Digital Twin Prototype — Combines modeled asset with operational/sensor/state data
- Operational Twin — Live system integration (not yet achieved)
Limitations
- Current prototypes use synthetic or simulated sensor data
- No live operational system integration has been verified
- A 3D viewer alone is not a digital twin
- Agent-based models use simplified behavioral assumptions
- IoT simulation does not represent real building systems