DOA estimation by fourth-order cumulants in unknown noise environments

Directions-of-arrival algorithms based on the cumulants of array data have been developed to suppress additive spatially correlated Gaussian noises. The authors propose a new cumulant-based method which is applicable in both Gaussian and non-Gaussian noise environments. This approach is derived on the basis of the fourth-order cumulants of the transformed array data by the discrete Fourier transform (DFT) and then relies on the statistical central limit theorem. Computer simulation results show that the proposed method is superior to the standard correlation-based multiple signal classification (MUSIC) technique and the cumulant-based MUSIC in both Gaussian and non-Gaussian noise environments.<>