Adaptive beamforming for moving targets using Genetic Algorithms and a CDMA reference signal

Adaptive Beamforming plays a key role in interference mitigation and target tracking. The Least Means Squares (LMS) algorithm has been successfully employed to accomplish this task given a reference signal. However, under severe reception conditions, LMS does not converge and locates the antenna beam on a wrong direction. Genetic Algorithms (GA) has shown to perform very well in global optimization tasks, even in blind DoA estimation under severe noise conditions. Then, this work investigates the use of GA for beamforming on a CDMA environment under different Signal-to-Noise Ratios (SNR), considering a reference signal is provided. Computational experiments on static targets showed that GA dramatically improved the robustness of beamforming to noise and converged faster when compared to LMS. For moving targets, due to its faster convergence, GA was able to track closely the target, while LMS algorithm barely reached it. On other hand, LMS performed better than GA on interference mitigation.