Zur Signalverarbeitung mit Statistiken höherer Ordnung
This habilitation thesis has to primary goals. First, the often mentioned prejudice against higher order statistics, that a great number of data is necessary for accurate parameter estimation, should be proven. In fact, we show by theoretical analysis of different applications for higher order statistics that this prejudice is not true in general. The main conclusion is that the use of higher order statistics has to be proven case-by-case. The second goal is to present a new concept - so-called modulated cumulants - for identification of linear time invariant system. We show by simulations that these concept leads to very accurate estimates compared to known methods.