Linear signal separation using polarization diversity

Abstract

A signal separator overcomes co-linear interference encountered in a wireless communication system base station by exploiting generally flat fading multipath characteristics of incoming signals. For this purpose, the invention employs diversity gained through the use of antenna array elements having different polarizations. An RF downconverter provides a signal separator with baseband, discrete-time digital samples of signal waveforms, received at one or more pairs of diverse polarization antenna elements. The signal separator weights and combines the baseband signals from the vertically and horizontally polarized received signals produced by the RF downconverter to estimate which signals emanate from which users. The signal separator uses polarization coefficients supplied by a coefficient calculator. To generate the polarization coefficients, the coefficient calculator is coupled to receive a set of channel fading coefficients from a channel estimator. Estimates of the channel fading coefficients may be derived using standard signal processing algorithms using the peak values and the known cross-correlation between training patterns. The coefficient calculator computes the polarization coefficients required by the signal separator by means of a coefficient matrix such that the output of the signal separator is equal to the user's information signal.