An investigation on semi-blind deconvolution by using nonlinear programming (Q1812617)

The semi-blind deconvolution problem (SBDP) [the authors, Sci. Sin., Ser. A 30, 1318-1332 (1987; Zbl 0642.94008); ibid., 1333-1344 (1987; Zbl 0642.94009)] is converted into a specific nonlinear programming problem. In this more general situation the relationship between the existence of a unique solution of the SBDP and the autocorrelation and crosscorrelation sequences of partial sample points is uncovered. As a result the robustness of the SBDP solution under noise situation is improved and the incompleteness of the SBDP solution method proposed in earlier papers of the authors (cited above) is reduced.