Separated estimation of wave parameters and spectral parameters by maximum likelihood

Parametric array processing methods are investigated. A conventional model of the sensor outputs is used: delayed signals are disturbed by sensor noise. Normal distributions of Fourier-transformed array outputs are applied for both, conditional and unconditional maximum likelihood estimates of the wave parameters, e.g. bearings and ranges, the spectral matrix of the signals and the spectral power of noise. We show that both methods allow to separate algorithmically the estimation of the wave parameters from that of the spectral parameters. We indicate iterative procedures for the estimates. Results of numerical experiments are presented which show that estimates of wave parameters and of spectral parameters can be significantly correlated for sources close together.