Sparse Array Optimization Based on Double-Deck ADMM for Near-Field MMW 3D Reconstruction of Human Bodies

In this letter, to address the problems of high cost and system complexity associated with uniform array antennas for human body inspection, a novel near-field sparse array manifold optimization algorithm is proposed based on a double-deck alternating direction method of multipliers (DD-ADMM) framework. Unlike conventional optimizations that requires manually adjusting the sparsity threshold of weight excitation, the proposed algorithm incorporates auxiliary variables to decouple the weight excitation from the objective function. Furthermore, an augmented Lagrange expression is constructed under the ADMM framework, so that the weight excitation can be automatically adjusted to obtain a closed-form solution for sparse elements. In comparison with the single-deck ADMM structure, the proposed algorithm mitigates the computational complexity associated with solving the joint minimization problem, which involves reducing the <tex-math notation="LaTeX">₁</tex-math> -Norm of weight excitation and minimizing the difference between peak sidelobe levels and auxiliary functions through the nested DD-ADMM framework. Additionally, the proposed algorithm handles sparse spatial-domain echo data, eliminating the need for interpolation and padding. Finally, the superiority of the algorithm is demonstrated through both simulated and raw MMW radar data for near-field human body imagery.