Model-Averaged Interference Rejection Combining

We present a novel algorithm for interference suppression when the number of interferers is unknown. The key idea is to parameterize the second-order statistics of the interference via a mixture of several low-rank models. Numerical examples illustrate the performance of the method.