Lattice implementation of Sum-squared Auto-correlation Minimization (SAM) channel shortener

The Sum-squared Autocorrelation Minimization (SAM) algorithm is one technique proposed for blind adaptation of the time-domain equalizer in multicarrier systems. The SAM cost depends on the effective channel autocorrelation, which will not be changed if any TEQ zeros are flipped over the unit circle. As a consequence, the SAM cost is multimodal, and different minima may yield very different shortening performance. In order to aid SAM converge to a suitable minimum, we impose a minimum phase constraint on the TEQ by means of the lattice structure and develop a lattice version of SAM with low computational complexity.