Lee, Gen-kwo; Gelfand, Saul B.; Fitz, Michael P. Bayesian techniques for blind deconvolution. (English) Zbl 0870.93046 IEEE Trans. Commun. 44, No. 7, 826-835 (1996). The paper presents a new algorithm for blind deconvolution in digital communication which performs joint input and channel estimation. The blind deconvolution problem is formulated as a nonlinear and non-Gaussian fixed-lag mean square error filtering problem, and the extended Bayesian filter is derived as a suboptimal recursive estimator.Extensive simulations characterizing the performance of the filter in severe intersymbol interference channels are presented. Reviewer: D.Jarušková (Praha) MSC: 93E11 Filtering in stochastic control theory 93C62 Digital control/observation systems Keywords:nonlinear non-Gaussian fixed-lag mean square error filtering; input and channel estimation; blind deconvolution; suboptimal recursive estimator PDFBibTeX XMLCite \textit{G.-k. Lee} et al., IEEE Trans. Commun. 44, No. 7, 826--835 (1996; Zbl 0870.93046) Full Text: DOI