Performance of a Reduced Complexity Algorithm for Diversity Combining

The paper introduces a new diversity algorithm that uses real valued weights to reduce the complexity of coherent combining. The approach starts with Selection Diversity but allows any signal branch to contribute to the output by a simple real valued scaling operation. Channel estimation is performed only on the branch with the strongest signal and the rest of the combining weights are then computed directly from the received signals themselves. This simplifies the weight computation and the resulting real-valued weights allows for combining to be performed at the analog or RF/IF stage, reducing the need for multiple receive chains required by digital combining. Simulation results reveal that the performance of the proposed algorithm is just 0.7 dB and 1.6 dB below optimum for 2 and 4 antennas respectively in independent Rayleigh fading channels.