Multidimensional nonlinear Schur parametrization of non-Gaussian stochastic signals. III: Low-complexity staircase solution (Q701944)

This is the third part of a paper on nonlinear Schur parametrization of non-Gaussian stochastic signals. It deals with finding a low-complexity solution of the above mentioned problem following from a multidimensional generalization of the two-indexed matrix extension problem. The results obtained allow for a considerable reduction in complexity of the Schur algorithms for stochastic sequences and orthogonal nolinear filters of the Volterra-Wiener class. For Part II, see ibid. 15, No. 3, 243--275 (2004; Zbl 1083.93029).