No abstract

A robust approach to estimation the intensity of a noisy signal with additive uncorrelated impulse interference is proposed. An occurrence of the additive uncorrelated impulse interference leads to increasing of the observed signal dispersion within some sections with impulse interference. Robustness of the intensity estimation is achieved by decreasing the influence of sections with impulse interference. A number of nonlinear filtering methods basing on lower envelope detection are developed: two-parameter recursive filter, dilation filter, clipping derivative filter and filters based on order statistics. Proposed approach was approbated by a numerical simulation. Numerical simulation is validated the efficiency of the proposed approach for estimation the intensity of a noisy signal with additive uncorrelated impulse interference at dynamic data mining and data stream mining.