To solve the problems of high computational complexity and unstable image quality inherent in the compressive sensing (CS) method, we propose a complex-valued fully convolutional neural network (CVFCNN)-based method for near-field enhanced millimeter-wave (MMW) three-dimensional (3-D) imaging. A generalized form of the complex parametric rectified linear unit (CPReLU) activation function with independent and learnable parameters is presented to improve the performance of CVFCNN. The CVFCNN structure is designed, and the formulas of the complex-valued back-propagation algorithm are derived in detail, in response to the lack of a machine learning library for a complex-valued neural network (CVNN). Compared with a real-valued fully convolutional neural network (RVFCNN), the proposed CVFCNN offers better performance while needing fewer parameters. In addition, it outperforms the CVFCNN that was used in radar imaging with different activation functions. Numerical simulations and experiments are provided to verify the efficacy of the proposed network, in comparison with state-of-the-art networks and the CS method for enhanced MMW imaging.

No abstract

A novel two-dimensional (2-D) compressive sensing (CS) based method is presented for near-field radar imaging. First, an accurate near-field approximation is proposed, based on which the circular wavefront curvature of spherical waves can be compensated by mapping the images to a rectified new grid. More importantly, the near-field approximation makes the two dimensions of the scattered data separable for the range and cross-range directions, which makes it possible to solve the 2-D reflectivity matrix for the image reconstruction directly. Then, a 2-D proximal subgradient algorithm for near-field radar imaging based on a fast iterative shrinkage/thresholding algorithm (FISTA) is introduced to resolve the memory usage and computation time issues. Simulation and experimental results are provided to demonstrate the performance of the proposed method with comparisons to the traditional Fourier-based method and to the conjugate gradient (CG) based method, which proves that the proposed method is an effective way to solve the near-field radar imaging problem.Index Terms-Near-field, radar imaging, spherical waves, compressive sensing (CS), FISTA. 0018-926X (c)

No abstract

ABSTRACT:With an aim to reducing manufacturing costs, in general and specifically to provide a solution to the thick laminate curing depth issue for composite materials, UV curing technology was combined with a fiber placement process to fabricate acrylate/glass-fiber composites. A novel layer-by-layer UV in situ curing method was employed in this article and interlaminar shear strength (ILSS) tests and SEM were used to evaluate the effect of processing parameters, including compaction force and UV exposure dose, on ILSS. The SEM images from short-beam strength test samples and the results of ILSS showed that the fibers' distribution was uniform in the cured matrix resin resulting from the compaction forces and that beneficially influenced the ILSS of the composite greatly. However, the matrix resin produced large shrinkage stresses when it reached a high degree of conversion (DC) in one-step, which resulted in poor interlaminar adhesion. In addition, the fast curing speed of UV on the composite resulted in poor wetting between fiber and resin, and accordingly resulted in lower ILSS. To overcome these problems and obtain high ILSS value composites, an optimized compaction force and UV exposure dose were determined experimentally.