On the interpolation of the weights of an OFDM-MMSE smart antenna for an infrastructure WLAN in arbitrary indoor environments

This work investigates the interpolation of the weights (multiplier coefficients) of a smart antenna optimum combiner serving multiple users through space division multiple access (SDMA) in a wireless LAN following the IEEE 802.11a or HIPERLAN/2 standards. The frequency dependence of the channel transfer matrix and the weights arc is first studied in indoor environments of different dimensions. A previously developed ray-tracing tool is used to predict multipath wireless channel parameters of interest. The performance of a minimum mean square error (MMSE) smart antenna in the low noise and high noise cases is then studied. The coherence bandwidth value is calculated from the frequency autocorrelation function and compared with the statistical estimation. The performance degradation of the SINR value due to interpolation is investigated for linear and FFT based interpolation when calculating every n^th subcarrier and interpolating the rest and with the variation of the coherence bandwidth of the indoor environment. Simulation results agree with the expected behaviour showing that for a small environment with a large coherence bandwidth few subcarriers are needed which reduces the complexity of calculating the weights at every subcarrier while for large indoor environments with a small coherence bandwidth, the weights need to be interpolated more frequently. The FFT based interpolation generally outperforms linear interpolation when the span between calculated subcarriers is less than 16.