Initial weight computation method with rapid convergence for adaptive array antennas

This paper presents a rapid initial weight computation method for adaptive array antennas. The proposed method estimates a correlation matrix and a correlation vector, where the correlation matrix is estimated using not only received samples with the desired signal, but also those before starting the transmission of the desired signal. The proposed method reduces the number of known symbols required for weight convergence. Furthermore, the combined use of the proposed method and the decision-feedback minimum mean square error combining method significantly improves weight convergence. Numerical results show that the combined method converges an array with sixteen antennas by using less than ten known symbols.