A New Improved Algorithm for Low Complexity Wideband Angle of Arrival Estimation

Abstract

The two fundamental problems of Test of Orthogonality of Projected Subspace (TOPS) based Angle of Arrival (AOA) estimation methods are computational complexity and spurious peaks. This work tries to explain the causes of these problems and applies a two-step modification to the Squared-TOPS (SQ-TOPS) procedure to tackle both problems. In the first step, the computational complexity is minimized via replacing the computationally-intensive eigenvalue decomposition (EVD) with a novel Covariance Matrix (CM) subsampling methodology to construct Projection Matrix (PM). Second, the problem of spurious peaks generation is addressed by establishing a new squared orthogonality test between the newly created PM and the transferred signal subspaces (after subspace projection). To justify the effectiveness of the suggested method, a numerical example along with intensive Monte Carlo simulations over different scenarios are provided where the proposed algorithm is systemically compared with its rival methods. The simulation results show the superiority of the proposed algorithm, projection matrix squared TOPS (PMS-TOPS), in terms of lower complexity, higher accuracy, less sensitivity to correlated sources, and low converges time in comparison to TOPS, SQ-TOPS, and weighted square TOPS (WS-TOPS) methods.