Near-field source localization based on higher-order statistics

A 3-D near-field source localization algorithm based on higher-order statistics using a small antenna array is presented in this paper. To suppress Gaussian color noises of unknown autocorrelation matrix in received array data, a specific fourth-order cumulant matrix instead of autocorrelation matrix is formed and an ESPRIT-like shift-in variance technique is used to estimate the phase differences among the antennas and ranges from the sources to the antennas simultaneously. Then a closed-form source location estimate is given by the solution of a set of linear equations. The proposed algorithm imposes no geometric constraint on the antenna array. Simulation results are provided to demonstrate the effectiveness and feasibility of this method.