Synthesis of algorithms for the tuning of adaptive systems in a nonstationary noise environment with pulsed and jumplike perturbation. I

The Markov theory of optimal filtering is used to derive signal processing algorithms for least-rmse adaptive systems in the presence of pulsed perturbations of the noise environment. A quasi-optimal algorithm in the Gaussian approximation is obtained for a finite number of pulsed perturbations. It is shown that, in the case of Poisson statistics of the perturbation sequence, the optimal algorithms can be derived directly from equations of the optimal filtering of discontinuous optimal processes.