New cumulant matrices for DOA estimation

We address in this paper the bearing estimation problem in the presence of spatially correlated Gaussian color noise. Two fourth-order cumulant matrices are presented. Beamspace data are processed to calculate the diagonal slice matrix and the contracted quadricovariance matrix, as a result the computation can be reduced by downsizing the cumulant matrices. The estimation performance, however, can be improved. Simulation results are presented to show that the beamspace matrices have higher resolution and lower root mean square errors than their equivalent methods in the element domain.