University of Minnesota Morris2013
B-Spline Parameter Estimation
Fitting differential equation models to noisy data using spline-based least squares
Summary
Applied B-spline least-squares fitting to estimate parameters of the Chua chaotic circuit from noisy experimental data.
Challenge
Estimate dynamical system parameters without requiring initial guesses — a common failure point in traditional methods.
Contribution
Implemented iterative B-spline fitting; developed rearranged matrix formulation reducing computation by ~10×.
Outcome
Demonstrated parameter recovery without initialization; rearrangement cut runtime from 30s to ~2s per iteration.
