Stabilization of spectral methods for the analysis of singular systems using piecewise constant basis functions (Q1901490)

The analysis of spectral methods applied to initial value problems for singular linear systems \[ Ex'(t)= Ax(t)+ Bu(t) \] is discussed. The methods can be sensitive to numerical errors, noisy input data, or incorrect initial conditions. Regularization methods are introduced in this paper in order to reduce the sensivity of the spectral methods and to stabilize them. These regularized methods can be rewritten in recursive form, leading to a class of discrete-time filters.