Minimum bias spatial filters for beamspace direction-of-arrival estimation

Spatial filtering is used in array signal processing to favor signals coming from a particular zone in space over other signals. In this paper, we propose the use of the minimum bias (MB) windows, originally devised for multitaper spectral analysis, as spatial preprocessing filters for direction-of-arrival (DOA) estimation. These windows (or tapers) are characterized by their orthonormality, simplicity, and most importantly their close behavior to the well known discrete prolate spheroidal sequence (DPSS) windows. A desired spatial bandwidth can be achieved by adding or removing tapers, without the need of recomputation. The paper introduces the MB tapers for beamspace processing and compares their DOA estimation performance, using noise-subspace eignestructure methods, to that of the DPSS beams. Further, the MB windows can be well approximated by simple trigonometric functions. These functions have expressions identical to that of the Discrete Sinusoidal Transform (DST). As numerous methods for fast computations of DST have been devised, the outputs of the approximated MB tapers can be computed efficiently.