A weight-vector LMS algorithm for adaptive beamforming

The conventional adaptive LMS algorithm has been utilized in array antenna beamforming to direct the radiated power towards the desired signal and null the multipath signals. In this paper, we present two new weight-vector adaptive LMS algorithms (WV-LMS) for minimum mean square error (MMSE) beamforming adaptive algorithm. Rather than using a fixed convergence parameter /spl mu/ in the conventional LMS algorithm, the two proposed algorithms exploit the useful information in the forward prediction of the weights vector. We used the forward prediction of the weights vector to dynamically change the convergence parameter /spl mu//sub n/. Both algorithms have been tested for beamforming using a narrow-band FM signal and have been shown to provide better MMSE performance than the conventional LMS.