Bounded-error parameter estimation: Noise models and recursive algorithms (Q1923018)

The paper deals with basic issues associated with the set membership approach to bounded noise parameter estimation problems: estimates consistency, structure of the experimental data and the a priori information about the noise. A relationship between various noise models (noise upper bounds) and the `size' of the set of parameter estimates being consistent with the data and the a priori information is established. Next, the efficient recursive estimation algorithms, producing parameter estimates that converge to the membership set, are presented for most commonly considered noise models. Also a recursive membership set parameter estimation method that is worst-case optimal is proposed.