Rekursive Methoden kleinster Fehlerquadrate unter Verwendung orthogonaler Transformationen

This paper presents identification methods which are especially suitable for an on-line and adaptive model estimation. The basis for these techniques are recursive variants of the least squares method and two extended least squares methods. All described methods apply orthogonal transformations as the most important and partly unrenounceable mathematical tools. These transformations have excellent numerical qualities and lead to very reliable and fast recursive identification methods