Blind Signal Estimation Using Structured Subspace Technique

Abstract

This brief addresses the blind signal estimation of a single input multiple output (SIMO) finite impulse response (FIR) system. The proposed method estimates directly the input signal with no need for the channel estimation stage or estimation of the second-order statistics. This is achieved by exploiting the Toeplitz structure inherent in the signal’s linear model, to build a quadratic cost function. Such methods are needed to lessen the latency and wave the error propagation phenomena. The simulation results are compared to other methods and are shown to highlight appealing advantages under adverse scenarios, specifically, short-burst communications and poor channel diversity.