Given Correlation Manifolds and their Application in Blind Channel Identification

This paper introduces the new objects named polynomial statistics. We discuss their properties and applicability to the problem of Blind Channel Identification which is a highly topical issue at the moment. This paper proposes a solution to this problem which is based on using polynomial representations of finite random sequence. Algorithms of Blind identification that was described in this paper use a factorization property of so-called given correlation manifolds.