Wall-Parameters Dependent Sparse MIMO Array Design for Ultra-Wideband TWRI

In through-the-wall radar imaging (TWRI), antenna array design is crucial for accurately reconstructing the dielectric profile of targets hidden behind walls. While deterministic or stochastic methods have been widely utilized to construct arrays in free space, challenges arise when dealing with the scenario of TWRI since the electromagnetic (EM) wave interacts complexly with wall materials. Moreover, many TWRI scenarios involve multilayered walls, a factor often overlooked in previous array designs. This work presents a wall-parameters dependent sparse multiple-input multiple-output (MIMO) array design approach for TWRI. First, the signal model for radar imaging through layered walls is presented, incorporating the transmission coefficients to represent wave propagation within the wall layers. Second, the averaged sidelobe level and the entropy of the Gram matrix are chosen as optimization criteria. The latter metric assesses the severity of sidelobe distribution in nontarget regions for a given array. Then, the covariance matrix adaptation evolution strategy (CMA-ES)-based multi-objective optimization, as its first reported implementation in TWRI, is employed to solve the array optimization problem. Finally, the finite-difference time-domain (FDTD) method-based numerical simulations and onsite experiments are provided to validate the effectiveness of the proposed array design strategy and show that the proposed optimized array achieves improved target reconstruction performance.