Dual Channel Matched Filtering and Space-Time Adaptive Processing

We propose a dual channel matched filtering system that addresses two key challenges in the practical implementation of a single channel matched filtering system: secondary data support and computational cost. We derive an exact expression of the dual channel normalized signal-to-interference plus noise ratio (SINR) in terms of random variables with known distributions and approximate expressions of the mean and variance of the normalized SINR. Using these approximate expressions, we demonstrated that the dual channel system requires half the secondary data to achieve nearly the same SINR performance as an equivalent single channel system. With the dual channel system, two reduced dimension weight vectors are used in place of the larger single channel weight vector, offering the potential reduction in computational cost. The key to the dual channel system is the efficient block diagonalization of the interference plus noise correlation matrix with a fixed transformation.