Transportation safety digital twins

Transportation safety digital twins

Integrating sensing, trajectories, microscopic simulation, safety metrics, and natural-language interfaces for transparent what-if analysis.

Overview

Why this research matters.

Digital twins can become more than animated simulations. My current work connects real-world observations, road-user tracking, SUMO simulation, surrogate safety measures, spatial risk maps, and comparative visualization in an integrated research environment.

The architecture distinguishes micro safety—individual road-user interactions and instantaneous conflict indicators—from macro safety, where risk accumulates by movement, intersection, corridor, and time. This makes it possible to explain both why risk occurs and where intervention may have the greatest value.

For tribal and rural corridors, the platform is being developed as a continuously synchronized decision-support system that can diagnose near-term safety states, test interventions before deployment, and translate sensor intelligence into driver-facing guidance through variable message signs. A local large-language-model pipeline also supports site screening and corridor-level safety audit reports without relying on cloud APIs.

Micro + macro

safety intelligence from interactions to corridor patterns

Local AI

analysis and reporting without a cloud API dependency

What-if

simulation of safety interventions before deployment

VMS-ready

path from detected risk to actionable driver guidance

Research questions

What I am working to understand.

  • How can observed movement be transferred into a calibrated simulation environment?
  • Which safety indicators provide a balanced view of scenario performance?
  • Can language-model interfaces improve access without obscuring analytical assumptions?

Visual notes

Figures, systems, and field context.

Resources

Publications, presentations, code, and data.

Public materials for this research direction are being prepared. Request research materials ↗