An Adaptive Update Method of Digital Twins for Physical Entities State Changes

Constructing real-time, high-precision depictions for changing physical entities is a current challenge in digital twin research. However, existing research on digital twins updates through real-time data is still insufficient in terms of update speed and effect. To address this, this paper proposes an adaptive updating method that can help digital twins quickly adjust according to changes in the physical entity. The method consists of two components: an Adaptive Training set Construction Algorithm and a dual-ended update mechanism. By choosing more representative new data, the Adaptive Training set Construction Algorithm can optimize the speed and effectiveness of digital twins updates; The dual-ended update mechanism can further optimize the update speed through refining the digital twins updating procedure. Comparative experiments with existing methods show that our method can help the digital twins learn more knowledge in a shorter time and better adapt to changes in the physical entity.