Partially blind antenna diversity equalization of a jammed fading channel

For wireless mobile communications systems operating on a jammed frequency-selective fading channel, the receiver performance can be improved by using the joint antenna diversity and equalization techniques to eliminate both time- and frequency-selective fades and jammers effects. The spatio-temporal processing techniques introduced previously use the reference part of the transmitted sequence only for channel/noise covariance estimation. We present a novel approach to joint channel identification/symbol estimation in the presence of colored noise based on a maximum likelihood criterion and exploiting both the training and unknown part of the transmitted sequence. This approach results in computationally efficient iterative algorithms which outperform existing techniques.