Robust adaptive beamforming based on sliding-window RLS algorithm

First, the role that diagonal-loading technique plays to resolve the mismatch and underestimation problems is explained, which are often encounterd in array signal processing as to estimate the covariance matrix. And then the drawback of RLS algorithm that the diagonal loading value attenuate rapidly is pointed out. Thus a new RLS algorithm using sliding rectangular window (RLS-SW) is presented to estimate the covariance matrix and invert it recursively. The simulation results shows that RLS-SW algorithm can maintain the diagonal loading value, which makes it more robust. Computational complexity of RLS-SW algorithm and other matrix inversion algorithm such as Gaussian elimination and RLS are analyzed and compared. The results show that the RLS-SW algorithm presented in this paper is efficient and robust enough to be applied to real-time adaptive beamforming.