Beamspace ML bearing estimation incorporating low-angle geometry

A problem in low-angle radar tracking, namely, bearing estimation in the presence of a strong specular multipath component that arrives within the beamwidth of the direct path signal, is studied. Three-dimensional beamspace domain maximum likelihood (3D-BDML) is a computationally simple ML bearing estimation algorithm applicable in this scenario which operates in a 3-D beamspace. A variation of 3D-BDML incorporating the multipath geometry as a priori information is presented. In symmetric 3D-BDML the pointing angle of the center beam is equal to the bisector angle between the direct path ray and the image ray, which may be estimated a priori given only the radar height and the target range. The effect of the inclusion of a priori information on the performance of 3D-BDML is analyzed in terms of the dependence on the relative phase difference between the direct and specular path signals, the sensitivity to error in the bisector angle estimate, and the results of operation when no specular multipath component is present in the data. In addition, computationally simple schemes for coherently incorporating multifrequency data into 3D-BDML are investigated     

Direction-finding with sparse rectangular dual-size spatialinvariance array

A novel sparse array geometry embedding two sizes of spatial invariances is presented for use with a new ESPRIT-based (estimation of signal parameters via rotational invariance techniques) algorithm for aperture extension. The half-wavelength invariance yields unambiguous but high-variance direction cosine estimates to disambiguate low-variance but cyclically ambiguous estimates from the larger invariance. With larger invariance at 60 half-wavelengths, resolution threshold for two closely spaced emitters is reduced by 50 dB and estimation error by 100-fold. Array design formulas are also presented