Design of Beamspace Transformation for Robust Dimension Reduction in Array Processing

Data reduction by employing a linear transformation to map the full dimension element-level data into a dimension-reduced beamspace has received considerable attention in array processing. A well-designed transformation offers significantly reduced computational complexity and greatly enhanced performance in subsequent processing. This paper addresses a key challenge in beamspace processing and presents a novel technique for beamspace transformation design. The problem is formulated as a constrained optimization problem, where the cost function evaluates the in-sector and out-of-sector performance criteria simultaneously, and then a numerical optimization paradigm is utilized to search for the optimal solution. The algorithm is flexible, robust and data independent, which performs well in harsh environments involving a small number of snapshots and highly correlated or coherent sources. The proposed framework can be further extended to fulfill multiple in-sector requirements, for instance, for both sensitive target detection and accurate direction-of-arrival (DOA) estimation in the context of radar applications. Simulation results validate the effectiveness of the proposed method and demonstrate the superior DOA estimation performance in the presence of strong sweeping interference.