Optimization-based State Estimation under Bounded Disturbances

TL;DR

This paper presents an optimization-based state estimation method for nonlinear systems with bounded disturbances, proving stability of the full information estimator and discussing its relation to existing approaches.

Contribution

It introduces a robust stability proof for the full information estimator in nonlinear systems and explores its connection to prior work.

Findings

FIE is robustly globally asymptotically stable for certain cost functions

Theoretical stability results are validated with a simple example

Discussion of relationships with existing state estimation methods

Abstract

This paper studies an optimization-based state estimation approach for discrete-time nonlinear systems under bounded process and measurement disturbances. We first introduce a full information estimator (FIE), which is given as a solution to minimize a cost function by using all the available measurements. Then, we prove that the FIE of an incrementally input/output-to-state stable system is robustly globally asymptotically stable under a certain class of cost functions. Moreover, the implications and relationships with related results in the literature are discussed. Finally, a simple example is included to illustrate the theoretical results.