Smart antenna performance and complexity for estimated Rician fading with correlated azimuth spread and K-factor

This work shows that the full performance permitted by the Rician fading radio channel at a smart antenna receiver with imperfect channel state information is achieved with maximal-ratio eigencombining (MREC) for much lower complexity (i.e., baseband power consumption) than with conventional maximal-ratio combining. Furthermore, it shows that unrealistic assumptions about the channel fading (Rayleigh, Rician with typical K value) produce not only several-dB performance estimation inaccuracies, but also up to 50% processing cost estimation inaccuracies. These results are obtained by deriving a new average (over the noise and fading) error probability expression for MREC, and then averaging it numerically over lognormal azimuth spread and K-factor distributions recently reported from measurements. A MREC adaptation criterion earlier proposed for Rayleigh fading is generalized to Rician fading and demonstrates an excellent dimension-reduction capability, for more power-efficient smart antennas.