A new look at observers for systems with measurement uncertainty

The work that follows is a continuation of the investigation of metric based nonlinear state transformations, but with applications to observers for plants with measurement uncertainty. In particular, these transformations are shown to be useful for linear systems, and appear to be highly effective when bounded sensor measurement uncertainty is present. It turns out that a nonlinear observer can be constructed based upon the existing Luenberger linear structure, i.e., Kalman filter. This nonlinear structure enables the estimated state reconstruction to become less sensitive to sensor uncertainty due to the transformation itself.