Knowledge modelling in signal parameter evaluation

The aim of this work is to describe the available knowledge about the investigated signal to enable evaluation of uncertainty of the estimated parameter at each stage of the measurement process. For this purpose, signal is represented as a point in a finite-dimensional number space. Then, each information item described by a set of appropriate relationships is transformed into a set of possible signals. Thus, set-theory operations are applicable to data fusion modelling. As examples, sets of possible signals are presented for cases when the signal mean power is known, when the signal harmonics content is available or when signal maximum and minimum values are known. Finally, the necessary condition is formulated for a given information item usability in a defined parameter estimation. A bunch of subsets of signal space dimensions is defined and utilised for this purpose.