Blind channel identification robust to order overestimation: a constant modulus approach

The problem of blind system identification (BSI) of a single-input multiple-output (SIMO) finite impulse response (FIR) channel is addressed. A plethora of methods and techniques for BSI have been proposed in the literature so far, including the subspace method. A difficulty with the subspace based identification methods is their sensitivity to misspecification of the channel order. We propose an identification algorithm robust to channel order overestimation. This algorithm is based on the minimization of a constant modulus (CM) in conjunction with subspace orthogonality criteria. Numerical simulations and investigations are presented to demonstrate the potential of the proposed algorithms.