An adaptive receiver with an antenna array for channels with correlated non-Gaussian interference and noise using the SAGE algorithm

An adaptive receiver that uses multiple antennas to provide diversity against fading is developed for operation in an impulsive noise environment. The noise components at each sensor are assumed to be correlated. A mixture of multivariate Gaussian distributions is used to model the noise. Using a training sequence, model parameters are estimated by iterative procedures derived from the expectation-maximization (EM) algorithm. These estimated parameters are then used in a likelihood ratio test to recover the transmitted signals. Simulations show that the proposed adaptive receiver is robust, and near-optimum performance can be achieved when sufficient training data is available.