Performance of Smart Antenna Arrays with Maximal-ratio Eigen-combining

Conventionally, maximum average signal-to-noise ratio beamforming (Max-ASNR BF) and maximal-ratio combining (MRC) are applied to antenna array systems. A recently-proposed combining approach based on the Karhunen-Loève Transform (KLT) followed by MRC of the strongest resultant signals, denoted herein as maximal-ratio eigen-beamforming (MREC), was claimed to be superior and flexible compared to Max-ASNR BF and MRC for correlated, imperfectly-estimated channels, but comprehensive analytical tools required to understand, evaluate and apply MREC to smart antenna arrays were not available. For Rayleigh fading and BPSK modulation, general MREC average error probability (AEP) expressions are derived herein for correlated channels which are imperfectly estimated. Particular cases of these formulas also describe Max-ASNR BF and MRC of the original untransformed received signals (MRCO). A criterion which allows flexible MREC tuning to the actual environment by trading-off detection performance versus complexity is proposed. The analytical results are applied for channel estimation using pilot-symbol-aided modulation (PSAM) and interpolation and they show that significant gains can be obtained with MREC-based smart antenna arrays in scenarios with significant correlation of the received signals, compared to MaxASNR BF as well as MRCO.