Real-Time DOA Estimation of Wideband Signals with Multidimensional Arrays via Signal Subspace Techniques.

2D Unitary ESPRIT is developed as a closed-form 2-D angle estimation algorithm for use in conjunction with a uniform rectangular array URA. In the final stage of the algorithm, the real and imaginary parts of the i-th eigenvalue of a matrix are one-to-one related to the respective direction cosines of the i-th source relative to the two major array axes. Reduced dimension beamspace implementations of 2D Unitary ESPRIT are developed along with adaptations for other array configurations including two orthogonal linear arrays. A novel approach to angle estimation in beamspace is also developed based on the observation that beamspace noise eigenvectors may be transformed to vectors in the element-space noise subspace. The transformed noise eigenvectors are bandpass, facilitating multirate processing involving modulation to baseband, filtering, and decimation. As these operations are linear, the Root-MUSIC ESPRIT based algorithm merely premultiples each beamspace noise signal eigenvector by a precomputed transformation matrix. Compared to previous beamspace implementations of Root-MUSIC or ESPRIT, this approach places no restrictions on the structure of the matrix beamformer. Extensions for the URA are developed based on Multidimensional multirate processing. AN