RDToC

Road Digital Twin ontology and configuration optimisation with semantic validation.

RDToC is a research prototype for designing road digital twin configurations under engineering and operational constraints. It extends the SmartRDT design workflow with a layered validation strategy: formal SHACL checks for structural completeness and fast rule-based checks for engineering compatibility during optimisation.

Research Motivation

Digital twin design for road infrastructure is a configuration problem: a usable system must combine sensors, algorithms, communication, storage, and compute deployment while satisfying performance, cost, reliability, latency, disruption, and sustainability requirements.

This repository explores how semantic validation can reduce invalid configurations and make optimisation outputs easier to audit.

Key Ideas

• ontology-based representation of road digital twin components
• NSGA-style multi-objective search over candidate configurations
• SHACL validation for configuration completeness
• fast engineering rule checks during evaluation
• comparison against baseline configuration strategies
• generation of Pareto fronts, reports, and paper-ready figures

Repository Structure

• main.py - compatible main workflow for running optimisation and reporting
• ontology_manager.py - ontology construction and SHACL-aware validation logic
• evaluation.py - objective evaluation and semantic fast checks
• optimization_core.py - optimisation core
• baseline_methods.py - baseline comparison methods
• shapes/min_shapes.ttl - minimal SHACL constraints
• RMTwin_Results_20260101_0017/ - stored experiment outputs and paper figures
• config.json - default experiment configuration

Validation Strategy

The project uses two layers of validation:

1. Formal validation: SHACL checks ensure that a candidate digital twin configuration contains required components, such as sensor, algorithm, storage, communication, and deployment.
2. Engineering fast checks: runtime rules catch practical incompatibilities, such as requiring GPU deployment for deep-learning algorithms or wireless communication for mobile sensing.

This split keeps the optimisation loop fast while preserving auditability.

Quick Start

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

python main.py --config config.json

Example Use Cases

• compare design options for road digital twin deployment
• audit whether generated configurations are structurally complete
• study trade-offs between cost, detection performance, disruption, latency, reliability, and carbon emissions
• produce visualisations for digital twin design analysis

Skills Demonstrated

• semantic modelling
• SHACL and ontology-inspired validation
• road infrastructure digital twins
• multi-objective optimisation
• technical evaluation and decision support
• research-grade Python workflow design

Status

Research prototype. This repository is intended to document and test design ideas rather than serve as a production system.