Accelerating Computation for Large-Scale Wide-Band RF Imaging

This work focuses on enhancing computational efficiency in wide-band radio frequency (RF) imaging. Specifically, we aim to efficiently generate multispectral RF images that contain information on range, angle, and frequency-selected responses, employing wide-band frequency-modulated continuous wave (FMCW) signals for probing. While wide-band RF imaging can improve imaging resolution, it significantly increases computational complexity due to the necessity of jointly processing signals across a wide frequency range. To address this challenge, we formulate the wide-band RF imaging as a non-convex optimization problem and integrate multi-dimensional fast Fourier transform (FFT) into the optimization process. We investigate both first-order and second-order proximal algorithms, demonstrating that both the gradient and the Hessian inverse can be efficiently calculated using FFT. Numerical simulations demonstrate significant improvements in computational efficiency. The time required for calculating the gradient and the Hessian inverse is reduced by factors of more than 100 and 10,000, respectively, resulting in a large reduction in overall processing time. Furthermore, our approach reliably generates multispectral RF images across a range of signal-to-noise ratios (SNRs).