Real-Time Adaptive Inter-Model Gap Identification for the Application of Digital Twins

Compared to mathematical models, digital twin models are more welcomed nowadays for research in autonomous systems due to their strong capabilities of imitating physical plants' behaviors. Empowered by physical simulation engines, digital twin systems have state propagation rules similar to physical systems, but there will always be gaps between models in the physical world and the digital world. To find a proper method to describe the difference between the physical model and the digital model for further model correction, this paper aims to develop a new online adaptive identifier that quantifies the inter-model difference at a fast speed. Terms with fractional orders are included to ensure fixed convergence time and bounded estimation error. A double-identifier structure is then established to ensure simultaneous gap approximation among the physical, theoretical and digital systems. To ensure the effectiveness of the proposed designs, the double-identifier structure is connected to both physical plants and the simulation environment for validation, where state estimation errors converge and inter-model gaps are identified. In the future, how to utilize the estimated inter-model gap to modify digital twin models such that the digital twins could imitate the physical system's behavior is a problem worthy of consideration.