Abstract:

The Regularized Iterative Adaptive Approach (IAA-R) is a powerful method for high-resolution Direction-of-Arrival (DOA) spectrum reconstruction. However, its performance degrades severely when the angular search is restricted to a narrow sector. This paper reveals that this degradation stems from a structural failure of IAA-R’s adaptive regularization in such ill-posed scenarios. To overcome this limitation, we propose the Dual Regularized IAA (DR-IAA), which augments the adaptive regularization of IAA-R with a novel, complementary term. This second regularizer is theoretically derived from the geometric incompleteness of the limited-scan array manifold, which we quantify as an “observation capability loss.” This dual-regularization framework determines the baseline regularization level directly from the array manifold’s geometric properties and the received signal power, ensuring algorithmic stability. Numerical simulations demonstrate that DR-IAA maintains high angular accuracy and provides a clean, stable DOA spectrum, significantly outperforming IAA-R in severely constrained scan-angle scenarios, including those with out-of-sector interference.