Iterative Gram-Schmidt orthonormalization for efficient parameter estimation

We present an efficient method for estimating nonlinearly entered parameters of a linear signal model corrupted by additive noise. The method uses the Gram-Schmidt orthonormalization procedure in combination with a number of iterations to de-bias and re-balance the coupling between non-orthogonal signal components efficiently. Projection interpretation is provided as rationale of the proposed iterative algorithm. Computer simulations are conducted to show the effectiveness of the algorithm.