Avtomat. i Telemekh., 2025 Issue 4,Pages 34–54(Mi at16530)
Stochastic Systems
A decomposition–autocompensation method for signal recognition based on the principles of continuity, invariance, multiplication, and ranking
with regular and irregular interferences
Abstract:
Considering the principles of continuity, invariance, multiplication, and ranking, we develop a novel optimal signal recognition method under essential a priori uncertainty, with application to real-time information-measuring systems. By assumption, in addition to random noise with an unknown distribution law but a given correlation matrix, the observation equation may contain a regular interference with an analytical finite-spectral representation and an irregular interference without any effective probabilistic model. The latter interference
can be described only by introducing some numerical characteristics and constraints confirmed
by the operation practice of a particular system. This method is invariant to the above in
terferences, does not require traditional state-space expansion, and ensures the decomposition
of the computational procedure. We analyze random and methodological errors as well as the
computational effect achieved. An illustrative example is given.
Fulltext:
PDF file (648 kB)
First page:
PDF file