A low cost noncoherent receiver with adaptive antenna combining for high speed wireless LANs

For the market success of high speed WLAN equipment cost and power consumption are of prime concern. One major hurdle is the equalizer which is required due to multipath fading. We study a low cost/low power hybrid receiver structure, which consists of a coherent segment for training sequence processing and a noncoherent segment for processing of the data field. The receiver uses an adaptive antenna array instead of an equalizer. Our main contributions are (i) a simple scheme for the complexity reduction (by 70%) of the combiner training engine, (ii) realistic performance results due to the use of selected channel measurement data sets instead of diversity channel models, and (iii) a hybrid training algorithm which can cope with imperfections of the combiner circuit.