Immunity of automatic modulation classification algotithms against inaccurate estimation of signal parameters

Automatic Modulation classification lays important role in the systems of electromagnetic spectrum monitoring. However performance of the algorithms is depreciated due to inaccurate estimation of signal parameters. The problem of unknown parameters is well documented in the literature however, authors usually propose improved algorithms in order to maximize probability of correct classification when one or more parameters are unknown. In this paper the aim is different. Immunity of selected algorithms against inaccurate preprocessing is investigated in order to assess how precise estimation of signals parameters should be.