A Tutorial on Bayesian Sequential Data Assimilation for Dynamic Data Driven Simulation

Data assimilation is an enabling technology for dynamic data driven simulation to support dynamic state estimation and online model calibration. It also plays an important role for enabling effective use of digital twin technology. This paper presents a tutorial on Bayesian sequential data assimilation for dynamic data driven simulation. A special interest is in data assimilation for discrete simulations. This tutorial targets the broader audience in the modeling and simulation community. It describes the sequential Bayesian filtering framework and introduces particle filters as a data assimilation method for general discrete simulation models. Several illustrative examples are provided to help beginners to grasp the concepts.