Back to search

Georgia Institute of Technology

Unsupervised Online Subspace Adaptation for Dynamic Data-Driven Digital Twins

Abstract

dc:description.abstract

Engineered systems are deployed in ever-changing operating conditions, contexts, and environments. The growing complexity of these systems makes it costly and difficult to physically test for all possible conditions and configurations before deployment. In response, engineering organizations are rushing to build digital twins to reduce costs and virtually evaluate system performance. A digital twin is a virtual construct that imitates a physical system and receives data feedback from its physical counterpart. These constant updates to the digital twin and calibration of twin parameters ensure that the twin accurately represents the physical twin's performance throughout its deployment. However, the vast majority of existing calibration methods are supervised techniques that require observation of a model's output to function. Existing work has explored the application of various unsupervised domain adaptation techniques on digital twins, but these implementations do not account for evolving and unknown domains. This thesis proposes an online domain adaptation method for digital twins by training models on a common latent space based on low-dimensional temporal structure within the model inputs. Subspaces of the model inputs in each domain can be found from time-delay embeddings of the data, then aligned by Procrustes analysis to yield a latent space that mitigates model extrapolation. Data can be assimilated sequentially to learn this embedding in an online manner. The proposed method is tested on a simulated aerobraking spacecraft to predict heat rate during atmospheric entry. Results demonstrate the method preserves the predictive accuracy of a digital twin encountering extrapolatory data, enabling an online approach to domain adaptation in digital twins that is currently not explored in the literature.

Degree

thesis:*
Name thesis:degree_name
Aerospace Engineering, PhD
Department thesis:degree_department
Aerospace Engineering
Grantor
Georgia Institute of Technology
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Wang, Paul
Advisor dc:contributor.advisor
  • Mavris, Dimitri N.
Committee members dc:contributor.committeemember
  • Kennedy, Graeme
  • Vamvoudakis, Kyriakos
  • Fischer, Olivia
  • Nguyen, Elisabeth
ORCID person:identifier.orcid

Rights

Language dc:language.iso
English

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/1853/81699

Chain of custody

source
Harvested from
Georgia Tech
Base URL
repository.gatech.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Wang, Paul. Unsupervised Online Subspace Adaptation for Dynamic Data-Driven Digital Twins. Georgia Institute of Technology, 2026. https://hdl.handle.net/1853/81699