A sliding window method to apply the Stokes space based polarization multiplexing technique in a time varying SOP scenario is proposed. The method uses the Stokes space representation of the signal samples SOP in order
to dynamically reverse the coupling between the two orthogonal polarizations. We show that the variables of the method, the number of samples used in each best fitting plane process and the number of samples after which the
inverse SOP rotation matrix needs to be updated, can be optimized in order to support endless time-varying SOP changes.