Fast adaptive microwave beamforming using array signal estimation

A new perturbation technique is proposed which enables adaptive beamforming (ABF) in microwave domain using a single-port beamformer. In this technique, for a system with L antennas, the weight vector is independently perturbed L times to obtain L correlated outputs. These outputs are then used to find a noisy approximation of the antenna array signal for the gradient vector estimation. The process of weight perturbations is performed faster than the Nyquist rate mainly to increase the temporal correlation of the consecutive antenna array signal samples and to lower the perturbation error. Performance of the proposed perturbation technique with the adaptive unconstrained least mean square (ULMS) algorithm is investigated for different channel scenarios. The ULMS algorithm with single-port beamformer converges in very high noise and interference levels and with a convergence speed close to that of the multi-port receiver. After convergence, both the single-port and the multiport beamformer algorithms achieve the same steady state signal to interference plus noise ratio (SINR) gain. Effects of weight quantization are also investigated for the single-port beamformer with the proposed perturbation technique.