Multi-Fidelity SINDy: Sparse Discovery of Nonlinear Dynamical Systems with Fidelity-Weighted Measurements
Researchers have developed a new method called Multi-Fidelity SINDy to discover nonlinear dynamical systems from data with varying levels of noise and fidelity. This approach extends the existing Sparse Identification of Nonlinear Dynamical Systems (SINDy) framework by incorporating Ensemble SINDy and Weak SINDy with a weighted regression derived from generalized least squares. The method has been validated on benchmark systems, including ordinary and partial differential equations, and demonstrated its effectiveness in forecasting the dynamics of a double pendulum system. The findings suggest that this multi-fidelity integration can improve model recovery by mitigating the impact of heteroscedastic noise, even utilizing lower-quality measurements. AI