On the Use of Newton-Based Methods in Fast Moving Horizon Estimation

This paper explores the application of Newton-based methods for solving the nonlinear optimization problem typical of moving horizon estimation (MHE). Specifically, we examine the performance of Newton, quasi-Newton, and Gauss-Newton methods, with a focus on their computational efficiency and convergence properties. Through simulation results, we illustrate the difference in performance across different scenarios, providing insight into the strengths and limitations of each method in the context of fast MHE.