A Knowledge-Grounded AI Engine for Digital Twins: Supporting Conservation Through Long-Term Monitoring
Year: 2026
Type: article
Abstract: This paper presents a knowledge-grounded Digital Twin framework designed to support the interpretation of long-term structural monitoring data. The approach combines AI reasoning models, project-specific knowledge retrieval, numerical observations, and spatial information within an interactive environment where specialists can explore monitoring data and formulate analytical requests. The framework is demonstrated through the Milan Cathedral case study, using ground-based measurements, satellite interferometry data, and a selected subset of technical reports covering more than two decades. A Cesium-based 3D environment, including gaussian splatting visualization, provides spatial access to the monitoring information. The application illustrates how specialists can use AI to retrieve relevant knowledge, guide computational analyses, and examine relationships between monitoring observations and previous engineering interpretations. New forms of interaction among specialists, AI, and monitoring information are illustrated and discussed, highlighting the complementary roles of accumulated knowledge, AI-supported analysis, and human expertise in the interpretation process.
Source: Remote Sensing
DOI: 10.3390/rs18223963
Institution: Politecnico di Milano
Language: English
Subfield: Geology
Field: Earth and Planetary Sciences
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,258
OpenAlex ID: W7235864280
