A random sets-based method for identifying fuzzy models

The nature of interference sources in signal processing is a key problem in many applied disciplines. These interferences are often modelled by random processes, although it has been shown that many models can be favourably modified when some of the uncertainty sources are treated as fuzzy experiments. Following this spirit, the objective of this paper is to build a mathematical model which explains a set of imprecise measurements taken on a physical system. Furthermore, it is assumed that the effect of unmodelled inputs to the system can be regarded as a random process, but the imprecision in the measures is better described by means of a possibility distribution.