The research and implementation of processing algorithm for a non-stationary signal with input sampled-data missing and intense impulse noise

The results of the signal real-time processing algorithm implementation for a non-stationary signal in a software development environment LabVIEW Platform are presented. It is proposed to use the method of singular spectral analysis - “Caterpillar”-SSA to identify trends and forecast changes in time series during the periods of intensive impulse noise. The fuzzy logic method is used to determine the anomalous values of the original time series data. Investigation and comparison of the considered forecasting methods and other methods of filtration is performed using computer simulation.